Software isn’t dying.
But it is being re-rated.

A weird thing is happening…

Public software has been absolutely smoked. Multiples have compressed. The market is acting like the old SaaS playbook is 100% broken. And yet — enterprise buyers still need software. They still don’t want to build everything in-house. And AI isn’t killing software; it’s changing what good software looks like.

That’s why this conversation with Morgan Livermore matters. He’s spent ~15 years investing across enterprise and infrastructure software at Accel, Geodesic, and Quiet Capital. Now he’s building SuperCruise — a fund focused on a very specific wedge: profitable, bootstrapped enterprise and infrastructure companies with a few million in revenue, very high growth, and founders who want aligned capital — not high risk rocket fuel that could blow the engine up.

Watch The Episode

Listen To Episode Here

This Weeks Vertical Titan: Morgan Livermore (Founder & GP Supercruise Capital)

Morgan is one of those investors who has seen a few software cycles up close.

He started in growth at Accel, where he looked at the kinds of businesses most people now view as canonical software outcomes — profitable, compounding, infrastructure-heavy companies like Atlassian-style businesses that didn’t need to light money on fire to matter. He then sharpened that lens at Geodesic and Quiet Capital.

What he’s building now with SuperCruise is a reaction to the last decade of venture excess.

His core belief is simple: there’s a large and undercapitalized class of software businesses doing $2M–$7M in ARR, growing fast, already showing signs of efficiency, and being pushed toward venture models that don’t actually fit them. Too much capital, taken too early, can create what he calls internal combustion — the company grows headcount faster than it grows product-market fit, and the whole thing becomes structurally misaligned.

So SuperCruise is designed around a different promise: be the first real institutional capital for founders who have already proven viability, help them reach escape velocity, and do it without forcing them into a growth-at-all-costs script that no longer maps to reality.

That worldview also explains why Morgan is so interesting right now.

He’s not just underwriting software companies. He’s underwriting the transition from classic SaaS to AI-shaped software economics — where margins, pricing, distribution, and defensibility all get re-written.

Watch The Epsiode

Listen To Episode Here

Sixteen ideas for building
the next era of software.

Not just for investors — for founders building the next generation of vertical software and vertical AI businesses.

Below are the biggest playbooks Morgan surfaced — not just for investors, but for founders building the next generation of vertical software and vertical AI businesses.

1) Don’t confuse capital with product-market fit

A lot of founders still operate as if fundraising is validation.

Morgan’s view is the opposite: prove the business can stand on its own first. Capital should supercharge a working engine, not become the engine.

When money gets used to paper over weak product fit or sloppy execution, it creates internal combustion.

Action item: Ask a brutal question: if you couldn’t raise for 24 months, would your current product and GTM still survive?

2) Optimize for escape velocity, not vanity growth

Morgan isn’t looking for companies growing “fast” in the abstract.

He’s looking for businesses growing fast enough to become undeniable — fast enough to become IPO-able, strategically valuable, or structurally durable.

That’s a much better lens than generic growth worship.

Action item: Replace “how do we grow faster?” with “what growth rate gets us to inevitability?”

3) SaaS isn’t dead — but old valuation math is

One of Morgan’s most important points: software may be oversold, even if it deserved a repricing.

The market is treating many software names as if the whole category is melting.

But enterprises still rely on these tools, and ripping them out is far harder than Twitter makes it sound.

Action item: Separate multiple compression from category collapse. They are not the same thing.

4) ARR is getting more slippery

For years, investors had a clean north star: ARR.

But ARR gets much fuzzier when pricing shifts from seats to usage, billing frequency changes, and revenue becomes tied to outcomes.

In AI, the unit of value is moving.

That means the unit of valuation eventually moves too.

Action item: If you’re an operator, define your real value metric now — before the market defines it for you.

5) The next moat might be… documentation

This was one of my favorite parts of the whole conversation.

Morgan shared an example of a company winning because its documentation is easily accessible to AI agents.

Not because it had the flashiest brand.

Not because it had the biggest sales team.

Because agents could actually find it, understand it, and implement it.

In a world where software gets discovered and evaluated by machines, documentation becomes distribution.

Action item: Treat docs like a GTM surface, not a support artifact.

A quick word from our incredible partner,
Parafin - the leading embedded capital partner.

• 4.8 star rating on Trustpilot, with over 500 reviews [source]

• 71% CSAT rating across all partners

• 84 NPS

• Serving top Fortune 500 companies, including Amazon, Walmart, and DoorDash

Check Out Parafin

6) AI agents change who the buyer is

Historically, software was sold to a human buyer through demos, decks, and relationships.

Increasingly, software will be selected, tested, or recommended by agents upstream of the human decision-maker.

That shifts the game toward accessibility, interoperability, and clarity.

Action item: Ask: can an agent understand what we do, how to use it, and why we’re differentiated?

7) The incumbent tax is real

Legacy software companies aren’t just fighting technical debt.

They’re fighting incentive debt.

If you run a large seat-based software business, moving to an AI-native or outcome-based model means attacking your own cash cow.

It might mean lower prices, fewer seats, and layoffs.

Founder-led challengers can make that leap far more easily than professionally managed incumbents.

Action item: If you’re a startup, go where incumbents are least willing to cannibalize themselves.

8) Outcome-based pricing is coming — slowly, then all at once

Seat-based pricing made perfect sense when software was basically licensed access to workflow.

But AI changes the customer question from “how many users?” to “what work got done?”

Morgan believes software increasingly gets paid on units of labor or business outcomes — tickets resolved, meetings booked, tasks completed.

Action item: Start testing pricing around completed work, not just access.

9) Deep domain insight matters more as code gets cheaper

If everyone can build faster, the bar for “we built software” drops dramatically.

That means defensibility shifts away from raw code and toward earned insight — the kind that comes from understanding the strange, specific, ugly details of a workflow better than anyone else.

Action item: Write down the 10 non-obvious workflow truths your team knows that a generic AI team does not. How can you build around that?

10) Vertical beachheads will compress

The old playbook was to dominate a niche for 7–8 years and then expand.

In AI, that window may shrink to 1–2 years.

Products can be built faster. Competitors arrive faster. Expansion matters sooner.

Action item: Plan adjacency earlier. Your wedge still matters — but your expansion clock starts sooner.

11) The compound startup is becoming the norm

Win a foothold, then sell more and more functions into the same customer.

That matters even more now because distribution is precious and context is compounding.

Once you own trust, workflow, and data exhaust, adjacent products get much easier to layer in.

Action item: Map your second and third products before your first one fully matures.

12) The long tail gets more interesting in AI

Historically, many software companies couldn’t profitably serve small customers because implementation and service costs were too high.

AI changes that.

Lower build costs and lower service costs mean more of the SMB and long-tail market becomes economically viable.

Action item: Re-open segments you previously ignored because ACV looked too small.

13) Free wedges still work

In AI and infrastructure, giving people a “taste” of the product still works extremely well.

Let users adopt something cheap or free, then monetize enterprise security, governance, scale, and support.

That playbook isn’t dead.

It may actually get stronger.

Action item: Identify what complement you can commoditize to pull adoption forward.

14) Gross margins may go down — and that’s okay

One of the most under-discussed AI realities: some of the best new software businesses may not look like classic 80%+ gross margin SaaS.

Tokens cost money. Inference costs money.

But if the market gets dramatically larger and the product does more labor, a slightly lower-margin business can still be much bigger and much better.

Action item: Don’t benchmark an AI-native business only against legacy SaaS margin expectations.

15) Taste becomes a moat when intelligence gets commoditized

If intelligence becomes cheap, then knowing what to build, how to package it, and how to make it feel trustworthy becomes more valuable.

In a world full of generic AI output, taste starts to matter more, not less.

Action item: Invest in product sharpness and user judgment, not just model capability.

16) Human trust still matters

This might be the most important counterweight in the whole conversation.

Yes, AI automates.

Yes, agents will mediate decisions.

But in vertical markets especially, buyers still want trusted partners.

They want someone who understands the workflow, the risk, the weird edge cases, and the consequences of getting it wrong.

In a world flooded with AI-generated slop, trust becomes easier to lose and more valuable to own.

Action item: Build your company so customers experience you as a partner, not just a vendor.

The old software era was built on seats.
The next one isn’t.

It will be built by founders who stay lean longer.

By companies that understand a domain better than generalist competitors.

By products that are legible not just to humans, but to agents.

By pricing models that track outcomes more closely than access.

And by teams that understand that trust, taste, and workflow insight may outlast almost every other moat.

So no — software isn’t dead. Bad software, bloated software, and mispriced software are getting re-rated. The next iconic vertical software companies will be born from this exact moment — when everyone else is busy confusing chaos for collapse.

Watch The Episode

Listen To Episode Here

Thanks for reading LINEAR. I reply to every email…
Have any questions, feedback, comments? Let us know, we work for you!

Shaped by lessons from decades of building and scaling 20+ SaaS platforms, it turns payments into a true growth lever. Read the guide to see how top platforms get it right. Read Now.

I think I’m finally starting to see where software goes from here…

Here’s the short version of the thesis: a small handful of horizontal AI Operating Systems are going to own the general workflows of consumers and businesses, and they’re going to become the most valuable companies in the world, by far. Everything underneath them gets redrawn — category by category, workflow by workflow.

At the same time, a new wave of Vertical AI Operating Systems is going to win the industries where context, depth, regulation, and services actually matter. The horizontal players won’t go all the way down into every market, they will be constantly fighting off crown seekers.

As that plays out, a massive amount of software becomes headless — still useful, but no longer the thing the customer opens. The value moves up to whoever owns the interface, the workflow, and eventually the work itself.

That’s the map. Could I be wrong on parts of it? Definitely. But the more I look at it, the more it lines up. And once you see it, it’s hard to unsee.

The way I think about it now, the new world breaks into seven distinct kingdoms:

The Seven Kingdoms, In Detail

Here’s how I’d break the new stack down. Seven kingdoms. Each one with a different shape, a different set of winners, and a very different game to play.

Kingdom I: General Consumer AI Operating Systems
OpenAI · Meta · Grok
GoT Parallel: Most like Dragonstone — strategic, well-defended, and built to command the serious campaigns.

This is not just another app people open. It’s the place they start and never leave. Where they ask questions, make decisions, create things, shop, research, plan, and increasingly take action.

Right now, ChatGPT is winning in consumer. It’s the closest thing we have to a default AI Operating System for the consumer internet — the place a growing share of people start when they want to ask, decide, create, research, plan, and increasingly take action.

But this throne isn’t settled. Meta has the distribution, the social graph, and the hardware ambitions to push their assistant in front of billions overnight. Grok has Elon, X, and a willingness to play by different rules. Google has Search, Android, YouTube, and Gemini sitting in front of literally billions of people every day, and they’re absolutely going to fight for the consumer surface too, as will Apple. All of them are coming for OpenAI, and the next few years are going to be a real fight.

Whoever ends up owning the consumer interface owns the relationship. This kingdom is going to be small in number and absurdly valuable.

Kingdom II: General Business AI Operating Systems
Anthropic · Google · Microsoft
GoT Parallel: Most like King’s Landing — whoever controls the capital controls attention, distribution, and power.

On the business side, Anthropic/Claude looks really well positioned to become the general business AI OS — the thing that starts to eat the horizontal business suite market the way Google and Microsoft did in prior eras. And don’t count Google and Microsoft out either. They’re going to fight like hell here. OpenAI is also pushing hard into the business layer with ChatGPT Enterprise, Teams, and a growing stack of work-focused features — they want this kingdom too, not just the consumer one.

Additionally, we are seeing all the horizontal productivity software try to win here - think ClickUp, Monday, Asana, etc.

In this new world, Claude (or whoever wins) isn’t a chatbot sitting next to your work. It is the workspace. It’s where teams write, analyze, summarize, retrieve knowledge, make decisions, and increasingly execute workflows across the general business stack.

Same playbook as the consumer side: integrate with everything, learn what matters, then absorb the most strategic categories natively.

Kingdom III: Headless Consumer Software
Bannermen to Kingdom I
GoT Parallel: Most like the Reach — productive, and valuable, but ultimately downstream from whoever rules the center.

Whoever ends up owning the majority of the consumer interface (ChatGPT, Grok, Meta, etc.), a lot of the consumer software you use today gets pushed underneath it. Some of it gets eaten entirely. A lot more of it survives, but as headless utilities sitting behind the AI OS — useful, even necessary, but no longer the front door.

These can still be good businesses. Think booking vacations, shopping, weddings, etc. The catch is that the platform above you can come for your category whenever it decides the margin pool is interesting enough. You’re useful. You’re just not safe.

Kingdom IV: Headless General Business Software
Vassals to Kingdom II
GoT Parallel: Most like the Westerlands — full of assets and infrastructure, but expected to serve the larger war machine above them.

Same dynamic, business edition. CRM, project management, internal docs, reporting, ticketing, payments, payroll, analytics, collaboration — a lot of these tools will keep mattering, but they’ll matter less as front-end destinations if the user lives inside Claude (or Google or Microsoft’s version of it).

Plenty of real billion-dollar companies will still get built here. They’ll just be built knowing the platform above them can integrate, learn, and eventually absorb. The ones that survive will be deeply technical, painful to rebuild, and embedded enough that the platform would rather work with them than redo their work.

Kingdom V: Vertical AI Operating Systems
Health · Law · Build · Coin
GoT Parallel: Most like the North — too large, too complex, and too specific for any southern ruler to fully control from afar.

This is the kingdom I’m most excited about, and the one I’m spending most of my time on.

On the surface, a Vertical AI OS will often look a lot like Claude or ChatGPT. Conversational. Intelligent. Feels like working alongside a really capable AI. Underneath the hood it’s a totally different product. It’s trained on industry data, fine-tuned for industry workflows, wrapped in compliance and human services, and built to compress complex industry work into a few clicks.

Why this works: the horizontal players have an endless roadmap just chasing the most valuable headless tools. They aren’t going to go all the way down into healthcare, legal, construction, insurance, financial ops, or logistics with the depth those markets need. That leaves real room for vertical winners with industry context, regulatory depth, services, GTM, and trust.

And unlike horizontal, vertical isn’t winner-take-all. Some markets get one monopoly. Some get two or three real winners. Some support four or five strong players for a long time. That structure creates room for a huge number of $10M–$1B companies, not just a handful of trillion-dollar outcomes.

Kingdom VI: Headless Vertical Software
Bannermen of Kingdom V
GoT Parallel: Most like the banner houses of the North — not glamorous, but absolutely essential to keeping the realm running.

Underneath every Vertical AI OS, there’s a stack of tools doing the unglamorous work — industry-specific compliance, scheduling, communications, data, embedded finance, point solutions for very specific parts of an industry.

I think this is where thousands of really good bootstrapped and lifestyle businesses get built. They’re not flashy end-user brands. They solve ugly, essential, highly specific problems for the vertical AI OS layer. And because the vertical players don’t have horizontal-scale resources, they’re far less likely to absorb you. This is a much safer place to live than headless horizontal.

Kingdom VII: AI Services
Trade routes to every land
GoT Parallel: Most like Braavos — the wealthy Free City built across islands and lagoons, lending its services to anyone who can pay.

The last kingdom is services — and I think it’s going to be way bigger and way more defensible than people expect.

The AI OS companies want software-esque margins. They want to own the interface, the intelligence, and the highest-leverage parts of the stack. What they don’t want to do is own every messy, exception-heavy, labor-intensive workflow in the real world. That creates massive room for services companies that bring human expertise and actual work completion to both horizontal and vertical AI systems.

These businesses should be much better than old-school services companies. Better margins. Better workflows. Better software leverage. Better operating discipline. But they’re still services in the most important sense: they help the customer actually get the work done when software alone isn’t enough. That’s a great place to be.

How To Survive & Thrive In The AI OS Era

Okay — if that’s the map, the more useful question is where should you actually try to play? Here’s how I’d think about it, kingdom by kingdom.

I. Most of us should probably avoid trying to build the top-level AI OS

If I’m being honest, I think most founders should probably stay away from trying to build one of the giant horizontal AI OS companies at this point.

That market is likely going to be won by a tiny number of players, and most of them will either be today’s leaders or tech giants that come in late with absurd advantages in capital, compute, talent, and distribution. There may be a few late entrants that matter. History usually leaves some room for that. But if you are going after this layer, you need to understand what game you are actually signing up for.

You are not really trying to build a normal startup.

You are trying to build one of the most important companies in the world.

And if you do go after it, you better be prepared to spend something that looks closer to $1T+ than most people are comfortable admitting.

That’s just not the lane I’d advise most people to pick.

II. Headless horizontal software can absolutely work — but it’s a dangerous place to live

I do think there will be real billion-dollar companies built underneath the horizontal AI OS layer.

The problem is not whether they can become valuable. The problem is how much extinction risk comes with the territory.

If you are building headless software or AI that serves the horizontal giants, you have to assume the platform can come for you at any point. If your category matters enough, they can absorb it. If your margin pool looks attractive enough, they can pressure it. If your functionality feels strategic enough, they can decide it should just be native.

So if you want to build here, you need to be more than useful.

You need to be necessary.

The best companies in this layer will own something technical, painful, deeply embedded, and annoying enough that the platform would rather work with them than rebuild them. That can become a fantastic business. But it is not a relaxing place to live. You need to know that going in.

III. Communication tools can still matter — but I don’t think this is some giant greenfield for new entrants

I do think a handful of communication tools remain large and important for a long time.

Video conferencing matters. Team communication matters. There are some categories like that which are just too core to how people work to disappear overnight.

But I don’t really look at this and think, “Wow, what a wide-open new frontier for startups.”

Maybe a few products show up late and plug into the AI OS systems better than what exists today. That can absolutely happen. But if you want to play here, I think you need to be a truly exceptional product person. The bar on product quality, taste, and user love has to be incredibly high. And even then, you are going to spend your life watching your back.

The dynamic here feels much more like Snapchat vs. Facebook than some peaceful little software niche.

You can build something people love. You can carve out real value. You can even get big.

But if the giants decide your product matters, you should assume they are coming.

IV. Vertical AI OS is where I’m spending my time

This is the bucket I’m most focused on.

This is where I’m spending my time, money, and my energy because I think this is where some of the most important and durable companies of the next decade will get built.

The opportunity, in my mind, is to build the Claude Cowork for a specific industry.

On the surface, the interface may look similar to the horizontal AI giants. It will feel conversational. It will feel intelligent. It will feel like you are working alongside an incredibly capable AI. But underneath the hood, it is doing something very different. It is built on the industry dataset, the industry language, the industry workflows, the industry edge cases, and the industry realities. And because of that, it can distill very complex industry work down into just a few clicks.

That’s the unlock.

The defensibility here comes from wrapping the Vertical AI OS in the things the horizontal players are much less likely to go deep on: human services layers, compliance and regulatory depth, embedded workflow knowledge, operational expertise, trust with the customer, industry GTM, industry brand, and the kind of real-world feedback loops you only get by living inside ONE industry.

That combination is what makes this market so interesting to me.

In a lot of verticals, software alone is not enough. A model alone will not be enough either. Even a beautiful interface alone is not enough. What matters is whether the product can actually help the customer get the work done in a way that is faster, more compliant, more trusted, and more complete than the alternatives.

I also like that this market probably will not be winner-take-all in the same way horizontal will be.

Some verticals may end up with one clear monopoly. Some may have two real winners. Some may support four or five strong options for a long time. Those are competitive dynamics I can live with. In fact, I prefer them.

Because that kind of market structure creates room for a lot more outcomes. It creates room for a bunch of billion-dollar companies, but it also creates room for an extraordinary number of businesses in the $10M to $999M range. That is a huge opportunity set. A lot of great companies will get built there. A lot of private equity firms will live and die here too.

V. There will be thousands of great businesses built as headless software tools (but most will be small)

I also think there will be thousands of great bootstrapped and lifestyle businesses built in this era.

My assumption is that most of these are not flashy end-user brands.

They are headless software tools that power the AI OS layer, whether horizontal or vertical.

They sit underneath the interface. They handle narrow but important jobs. They solve ugly, essential, highly specific problems. And they become part of the stack that the AI OS relies on in order to actually function well.

Some will be data providers. Some will handle payments. Some will manage compliance. Some will power execution in very specific parts of a workflow. Some will connect fragmented systems that the AI OS needs in order to actually get work done.

A lot of these will become very good businesses.

In many cases, they will be built by founders who understand one annoying part of the stack better than anyone else and quietly turn that into a durable, profitable company.

VI. AI services will be huge — and much more defensible than people think

I also think there will be an explosion of massive AI services companies.

Why?

Because the AI OS companies will want software-esque margins. They will want to own the interface, the intelligence layer, and the highest-leverage parts of the stack. What they will not want to do is own every messy, labor-intensive, exception-heavy workflow in the real world. That creates a lot of room for services companies that bring human expertise and work completion to both horizontal and vertical AI systems.

These businesses should be much better than old-school services businesses.

Better margins. Better workflows. Better software leverage. Better labor models. Better operating discipline.

But they will still be services businesses in the sense that they help the customer actually get the work done when software alone is not enough.

That is a very good place to be. A defensible place to be. And a very large place to be.

VC’s probably will spend a fair amount of time here.

VII. The companies that move too slowly are in serious danger

The harsh reality is that a lot of companies are just not going to make this transition.

Some won’t see it clearly enough. Some will see it, but move too slowly. Some will keep pretending that adding a few AI features to an old product is the same thing as adapting to a new era.

It isn’t.

A lot of SaaS companies are going to find out the hard way that their position in the stack was weaker than they thought. A lot of services companies are going to realize too late that customers no longer want the old model of labor delivery. A lot of businesses are going to get caught in the middle: not differentiated enough to win, not lean enough to survive, and not fast enough to reposition.

That culling feels inevitable to me.

To Wrap It Up…

This isn’t me about telling you what type of company to build…

It’s about recognizing that the stack is getting redrawn in real time.

A few AI OS companies will sit at the top.
A huge amount of software will get pushed down, abstracted away, or turned headless.

Some companies will become core infrastructure.
Some will become vertical systems.
Some will become service layers.

And a lot of companies will realize too late that the value already moved.

That’s the real question now:

Or are you the company that wakes up one day and realizes it was never as defensible as it thought?

That’s why this moment matters.

Because this isn’t some slow transition that plays out over 20 years.

It’s happening now.

And the people who understand where the value is moving — before everyone else does — are the ones who get to matter in what comes next.

Godspeed.

Hear ye, hear ye—

Unto the brave founders, the bold builders, and all who would dare carve a kingdom from the new software realm — greetings.

If thou hast journeyed thus far down this chronicle, know that the gates are quietly opened. We do hereby summon thee to the

Vertical Software Summit in Miami on November 4th and 5th, 2026.

The formidable lords of every great vertical realm — founders and sovereigns of the mightiest houses — shall gather to share the chronicles of their conquests, their scars, and their unspoken strategies.

It shall be the vertical AI gathering of the year.
None like it shall come again.

Come, then. Take thy place among the worthy.

Sealed this day,

The Council of Vertical

Alex Jekowsky built one of the most defensible companies in vertical SaaS by doing what most investors told him not to do: he went all-in on hardware.

What if the best moat in software isn’t software at all...?

Watch The Episode

Listen To Episode Here

This Weeks Titan: Alex Jekowsky (Founder & CEO of Cents) & His Laundromat SaaS Empire

Alex Jekowsky is the co-founder and CEO of Cents, and he’s just pulled off one of the most impressive rounds I’ve seen in vertical SaaS this year: A $140 million Series C.

The numbers tell the story: Cents today powers over 4,500 locations, processes over $1 billion in payments annually, and operates in nearly 1 in 6 laundromats across the United States. In an industry that most tech investors wouldn’t even look at, Alex has captured massive market share by doing something most founders are afraid to do: betting on hardware, going deep on operations, and building for the long game.

The journey started after Alex sold his previous company, Ulyngo, in the higher ed tech space back in 2019. During his earnout period, he started looking into buying laundromats as a side investment and discovered something wild: 70-80% of the laundry industry was still coin-operated.

While dozens of venture-backed consumer laundry apps were trying to be the “Uber for laundry” and failing spectacularly, Alex saw the real opportunity. Those delivery companies were just bringing orders to laundromats anyway—the operators were the backbone with a 99% success rate. Alex made a different bet entirely: build for the operators, not the consumers.

What I admire most about Alex’s journey is the conviction to make contrarian bets when everyone said he was wrong. Going deep on laundromats when consumer apps were hot. Betting on hardware when every VC said it would kill his margins. Burning over $1 million per month in the early days—double what software-only peers were spending—because he knew hardware was the foundation.

The result? A $140 million war chest, dominant market share, and a business model that gets stronger with every customer. The hardware moat that VCs laughed at five years ago is now what makes Cents nearly impossible to disrupt in the AI era. That’s vertical SaaS done right.

Watch The Episode

Listen To Episode Here

Playbook #1: Should You Build Hardware? Alex’s Selection Framework

The answer depends on whether hardware solves the core problem in your vertical. For Alex, hardware wasn’t optional—it was the only way to get laundromats off coins and onto digital payments. But building hardware as a startup is expensive and full of landmines. Here’s how Alex did it without destroying his burn rate.

Bring Hardware Development In-House

Almost all founders outsource hardware development to contract manufacturers or agencies. You’ll get taken for a ride, find out about problems too late, and burn capital you can’t afford. Alex brought hardware development fully in-house from day one.

But how do you hire world-class hardware engineers when you’re pre-revenue? Alex’s solution: hire recent graduates from top hardware companies, give them meaningful equity, and let them build from scratch. He found Cents’s head of hardware through a mutual friend—a recent Apple grad. Tiny contractor fee, meaningful equity stake.

Here’s the recruiting framework: Don’t try to poach senior engineers from Apple or Google. Instead, make friends with those senior engineers and ask them to refer junior talent they’ve worked with. Junior engineers at great companies are well-trained, hungry for equity, and want the zero-to-one experience. Alex used this to build his entire early hardware team.

Accept the 2x Burn (And Fundraise Accordingly)

Hardware will burn roughly double what software-only peers burn. Cents was burning over $1 million monthly while comparable startups burned $400K-$600K. Why? Industrial design, supply chain infrastructure, QA testing, inventory management, PCI compliance, field testing—these costs are unavoidable.

But not all burn is equal. Alex framed hardware investment as the foundation of everything else. Without controlling the payment terminal, you can’t capture transaction data. Without transaction data, you can’t build the operating system. Without the operating system, you’re just another POS system that gets ripped out.

Investors who understood deep vertical SaaS got it. Investors looking for software-only multiples passed. That’s fine—you want investors who understand your strategy.

Use M&A to Accelerate (Not Pivot)

By Series B, Alex had deployed thousands of devices. But hardware development cycles are long, and the market opportunity was NOW. So after 16 months of relationship-building, he acquired Laundry Works in January 2023—a company with proven hardware in thousands of locations.

This wasn’t a pivot. It was strategic acceleration to de-risk execution on a pre-existing roadmap. Building equivalent scale organically would have taken 36 months. He didn’t have that time.

Alex’s M&A framework:

• Does this accelerate our existing roadmap or distract from it?

• What would it cost to build this organically in time and capital?

• Can we integrate quickly without a multi-year nightmare?

• Does this de-risk a critical part of our strategy?

Most investors push back on early M&A. Alex countered: “We’re not changing strategy—we’re accelerating what we already sold you on.”

Hardware as Your AI-Era Moat

Five years ago, VCs told Alex hardware would hurt his valuation. Today, in the AI era where software features can be replicated instantly, hardware is physical defensibility. You can’t ChatGPT your way to controlling payment terminals or building supply chains.

Software is infinitely replicable. Hardware is not. Even if competitors build identical software, they can’t easily steal customers. Switching requires ripping out terminals FROM EVERY Laundry machine, installing new hardware, retraining staff, migrating data, and dealing with downtime. That’s massive friction that software-only products don’t have.

A quick word from our incredible partner,
Parafin - the leading embedded capital partner.

• 4.8 star rating on Trustpilot, with over 500 reviews [source]

• 71% CSAT rating across all partners

• 84 NPS

• Serving top Fortune 500 companies, including Amazon, Walmart, and DoorDash

Check Out Parafin

Playbook #2: The Distributor-Led GTM That Scales

Early on, Cents tried the traditional SaaS playbook: hire salespeople, run ads, do outbound. It worked for initial traction but didn’t scale. Laundromat owners aren’t on LinkedIn or browsing SaaS directories. They’re busy operators running cash businesses.

Then Alex had a revelation: In hyper-verticalized markets, the person your customer already buys from is their most trusted advisor. For laundromats, that’s the equipment distributor. These distributors have decades-long relationships with store owners. They sold them their equipment, service it when it breaks, provide financing, give business advice. They’re consultants, advisors, friends. When a distributor recommends something, operators act on it.

Partner With the Channel, Don’t Disrupt It

Most tech companies see distributors as legacy middlemen to disrupt. Alex made them his core GTM strategy. Today, Cents’s hardware is often sold through laundry equipment distributors. The distributors are a massive sales force. Cents provides product, training, marketing materials, and support.

This gives you instant access to every operator without building a massive sales team. It leverages trust distributors spent decades building. It aligns incentives—distributors make money when operators upgrade. It creates geographic density organically.

But you can’t just show up with a product and expect distributors to care. They’ve seen startups come and go. You have to prove you’re committed for the long haul.

So Alex did something extreme: his team attended 60 distributor trade shows in 2 months. Constantly on the road, setting up booths, doing demos, answering questions, shaking hands. The goal was proving Cents wasn’t some fly-by-night tech company.

It worked. Distributors started seeing them everywhere. Today, the distributor network is one of Cents’s primary growth engines, scaling in a way direct sales never could.

The CAC vs. Channel Economics Trade-Off

Yes, you share economics with distributors. They take margin on hardware sales and sometimes small rev share on software or payments. But here’s what Alex understood: the alternative to paying distributor margins is a brutal CAC.

If you’re doing direct sales in fragmented SMB markets, your CAC might be $5K-$15K per customer. Plus you’re paying sales rep salaries, benefits, tools, travel. With distributors, you’re essentially outsourcing that entire function to people who already have the relationships and trust. Your “CAC” becomes the margin you share, but it’s variable (only when deals close) and you get instant scale.

This is why Cents has a larger Customer Success team than Sales team. The distributors handle acquisition. CS handles retention and expansion across multiple locations. In low-churn verticals, growth comes from expansion, not just new logos.

Playbook #3: The Fundraising Framework That Eliminates Negotiation

Alex’s Series C fundraising strategy is one of the smartest approaches I’ve heard. Most fundraising turns into negotiation about valuation. You justify why you’re worth X, investors try to knock you to Y. Alex flipped the script.

The “Non-Negotiable” Valuation Model

Instead of negotiating, Alex did the math FOR investors. He built a sum-of-the-parts valuation model projecting monthly out to 2031. Cents has three revenue streams: Software (SaaS), Payments (transaction processing), Hardware (equipment sales). He applied market multiples to each, added them up for total enterprise value, then built a Share Price Calculator showing exactly what investors get at a given price and their expected 5-7 year return. He also included every possible company comp in the model to drive the average multiple for each category.

Then he picked a price delivering a 3-5x return for growth equity investors—the standard expectation for that asset class.

When he walked into investor meetings, he walked them through the math. “Here’s our model. Here’s the market data. Here’s the formula. Here’s the price. We’re not negotiating valuation—we’re looking for the right partner.”

He eliminated negotiation and focused on fit. Investors either got it or didn’t. The ones who got it moved fast.

Why This Works

You look sophisticated. Investors respect founders who understand their numbers. You set the tone—you’re not begging for capital, you’re offering a calculated opportunity. You save time—no months of back-and-forth on valuation.

Pro tip: This only works if your numbers are rock-solid. Alex learned this the hard way. Early on, he reported “cARR” (Contracted Annual Revenue). Later, he switched to “Recognized GAAP SaaS Revenue”—the most conservative metric possible.

The Employee Tender Strategy

Here’s a killer recruiting hack: As part of the Series C, Cents ran an employee tender where employees with 2+ years tenure could sell up to 25% of their vested equity. This was $30 million of the $140 million round.

Why? Because Alex competes for talent with AI startups and Big Tech offering $500K+ packages. By offering early liquidity, he can tell recruits: “You don’t have to wait 10 years for life-changing money. If we hit milestones, you can take chips off the table in 2-3 years.”

It rewards early team members who took risk. And it’s a massive recruiting tool when competing with companies that offer higher cash comp but no near-term liquidity.

To Close It Out…

Alex’s journey with Cents is one of the best examples of contrarian thinking in vSaaS. He went into an unsexy vertical. He bet on hardware when everyone said don’t. He partnered with distributors instead of disrupting them.

The result? $140M Series C, 1-in-6 market share, over $1 billion in annual payment processing, and a business that’s nearly impossible to disrupt.

Here’s what is worth taking away:

Hardware can be your moat. If customers have a physical pain point, solve it with hardware and own the value chain.

Partner with the channel, don’t disrupt it. In hyper-verticalized markets, the existing trusted advisors are your best GTM.

Do the math for investors. In later-stage rounds, stop selling vision and start showing math.

You’ve got this. Now back to building!!!

Watch The Episode

Listen To Episode Here

Thanks for reading LINEAR. I reply to every email…
Have any questions, feedback, comments? Let us know, we work for you!

Shaped by lessons from decades of building and scaling 20+ SaaS platforms, it turns payments into a true growth lever. Read the guide to see how top platforms get it right. Read Now.

This week I went back and looked at one of my favorite old vertical SaaS visuals: the Toast graphic from their S-1.

It still hits…

Not because it’s polished. Not because it says “platform.” And not because it has a bunch of product boxes hanging off a restaurant. It hits because it captures something most software companies miss:

A vertical business is not a collection of features. It’s a collection of workflows.

That’s what Toast understood.

They didn’t just map restaurant software. They mapped the restaurant itself.

They mapped the way guests move through the space. The way staff move through shifts. The way money moves through the register. The way orders move through the kitchen. The way labor, inventory, reservations, reporting, loyalty, payroll, and payments all connect into one living system.

That’s why the visual has held up so well.

It wasn’t really a product diagram.
It was an operating diagram.

And I think that matters even more now than it did back then, because if we’re being honest, most vertical SaaS founders are still doing the old exercise. They’re still mapping the business as software categories. POS. Payroll. Scheduling. CRM. Analytics. Reporting. Maybe a little fintech layered in.

But AI changes the exercise.

Not a little.
Completely.

Because old software mostly helped you record work.
AI-native software can increasingly help you do the work.

That means the map itself has to change…

Toast didn’t just sell restaurant software.
They mapped the whole restaurant.

The easiest way to misunderstand Toast is to say they built restaurant POS software.

That’s true in the same way that saying Amazon is an online bookstore is true.

It’s technically correct and strategically useless.

What Toast really did was much more important: they took a fragmented, chaotic, deeply operational vertical and mapped it as a connected system. They understood that a restaurant is not just a place where transactions happen. It’s a place where dozens of small workflows collide all day long.

A table gets seated. An order gets modified. A ticket backs up. A shift changes. A supplier is late. A payment fails. A line cook doesn’t show up. A guest leaves a bad review. A manager comps a meal. A server picks up another section. A catering lead comes in. Inventory drops faster than expected. Payroll closes at the end of the week whether you’re ready or not.

That is the business.

And what made Toast so strong is that they didn’t look at those as isolated problems. They looked at them as parts of the same operating system.

That’s such a big lesson for vertical SaaS.

The best companies in our world do not start by asking, “What software category can we sell into this industry?”

They start by asking, “How does this business actually run?”

That sounds obvious, but it’s not how most people build.

Most people start with a wedge that looks like a screen. Toast’s insight was that the real wedge was the workflow. Once you understand the workflow, you can keep expanding through it. Once you sit close enough to the operational heartbeat, you get access to more use cases, more data, more trust, and eventually more monetization opportunities.

That’s why companies like Toast become more than software companies. They become the infrastructure layer for the vertical.

And, of course, once you really own the workflow, you don’t just sell software anymore. You start participating in the money movement underneath it. Payments. Capital. Payroll. Financial services. That’s where vertical SaaS got really interesting.

So when I look at that old Toast visual, what I see is not a company that won because it had a better interface.

I see a company that won because it had a better map.

And here’s the question I keep coming back to:

If we were doing that mapping exercise today, in a world where AI can actually interpret, route, recommend, and increasingly act, would we draw the same map?

I don’t think we would.

Not even close…

Old Vertical Software Mapped Workflows
But AI Should Map the Work

This is the real idea.

For the last twenty years or so, most vertical software was built around the same basic logic. Someone opens a screen, enters information, saves it, and then that information sits in a database until someone else needs it later. Maybe it gets turned into a report. Maybe it triggers a basic rule. Maybe it just lives there forever.

That model created an enormous amount of value. It took industries that were running on paper, memory, spreadsheets, and manual coordination and gave them structure. That mattered. It still matters.

But it also created a generation of software that mostly recorded the output of human labor instead of reducing the need for that labor in the first place.

That’s the subtle but important distinction.

The software became very good at storing decisions after a human made them. It was much less good at helping make the decision, moving the decision forward, or resolving the ambiguity that created the work in the first place.

That’s why so much software still feels passive.

It stores. It displays. It reports.

But it doesn’t really orchestrate.

AI changes that.

Now software can take in messy information, figure out what it means, structure it, compare it against context, recommend a next step, and in some cases just do the next step. That’s not a minor product improvement. That’s a different category of software behavior.

Which means if you’re a founder mapping a vertical today, you shouldn’t just be asking where data gets entered.

You should be asking where work begins.

That’s a much more interesting question.

Where does the actual workflow start? Is it an inbound phone call? A voicemail? An email? A PDF invoice? A reservation change? A catering request? A guest complaint? A no-show? A chargeback? A schedule swap? A missing item from a supplier? A manager text at 6:12 a.m. because somebody called out?

That’s the map now.

Not modules. Not menus. Not navigation tabs.

Work.

And if you really follow that thread, the whole architecture of the business starts to look different.

A restaurant, for example, is no longer just POS, payroll, scheduling, inventory, and marketing.

It becomes a chain of workflows.

Demand comes in from all over the place. Some of it is structured, some of it is messy. That demand has to be interpreted, converted, routed, fulfilled, measured, and followed up on. Labor has to flex around it. Inventory has to support it. Payments have to capture it. Exceptions have to get resolved. Managers have to step in only when the situation becomes ambiguous, high-risk, or economically meaningful.

That is a far more AI-native way to see the business.

And this is exactly why I think the “messy inbox” idea David Haber penned is such a strong thinking framework for vertical AI.

Most industries still have some version of this problem: too much unstructured stuff coming in at the top of the funnel, too many humans needed to interpret it, and too many downstream workflows that don’t start until somebody manually turns chaos into structure.

That’s where AI gets really powerful.

Not when it’s bolted onto the end of a workflow as a chat feature.

But when it sits at the beginning of the workflow and turns ambiguity into action.

That’s a very different thing.

And it’s why, if I were redrawing the Toast map for today, I wouldn’t organize it around product modules. I’d organize it around moments of operational gravity.

Where does demand enter the system?

Where does labor get burned?

Where do handoffs happen?

Where does money leak?

Where does a human still spend time interpreting something that software can now reasonably understand?

And maybe most importantly: where should the human still stay in the loop?

Because I don’t think the right AI map is one where everything gets automated. That’s fantasy. The right AI map is one where the software handles the clean, repetitive, high-frequency parts of the workflow and then pushes the weird, high-stakes, low-confidence moments into an exception layer.

That, to me, is the future UI of vertical SaaS.

Not the dashboard.

The exception queue.

A manager does not want another screen full of analytics. They want to know what needs attention right now, what was already handled, what is blocked, and what matters financially.

That’s the new operating system.

And honestly, once you see it that way, it becomes hard to unsee.

Old vertical SaaS mapped where information got entered.

Vertical AI should map where decisions get made, where work gets stuck, and where software can now carry more of the load.

That’s the exercise.

That’s the opportunity.

And I think that’s where the next great vertical companies are going to come from.

How To Re-Do The Exercise in the Age of AI

Part of what pushed me to write this came out of a conversation I had with David Haber, a GP at a16z.

We were talking about old vertical software maps, why some of them were so powerful, and why they now feel incomplete. That conversation sharpened something for me that I had been circling for a while: in the AI era, it is not enough to map the workflows.

You have to map the work.

That sounds like a small distinction, but I think it changes everything.

A workflow is the clean version of the business. It is how the business gets described in a deck, an implementation plan, or a software menu. Orders flow here. Payroll goes there. Scheduling sits over here. Reservations feed into table management. Inventory connects to purchasing. Reporting rolls up on the back end.

That is useful. But it is also sanitized.

Work is messier than workflow.

Work is the manager reading a long email from an angry guest and deciding whether to comp a meal.
Work is someone listening to a voicemail about a catering request and figuring out whether it is real, urgent, and worth following up on.
Work is a back-office employee opening an invoice PDF, checking whether it matches expectations, routing it to the right person, and cleaning up the edge cases.
Work is somebody noticing that a reservation change is going to affect table flow, staffing pressure, and probably ticket times.
Work is somebody acting as the human bridge between a messy real-world event and a structured software system.

That is the piece most old software never really touched.

The old software map showed you where information ended up.

It did not show you how much labor was required to get the information from here to there.

That is why I think the first principle for re-doing the Toast exercise in the age of AI is this:

Don’t just map the workflows — map the work required to keep the workflows moving.

That means asking a different set of questions.

Not just: what are the modules?

But: where is labor being spent on interpretation? Where is labor being spent on coordination? Where is labor being spent on routing, exception handling, follow-up, checking, clarifying, escalating, and cleaning up?

Because that is where AI enters.

If you were doing this for restaurants, you would still want to understand the big functional areas. You would still care about reservations, ordering, kitchen operations, labor, payments, and guest engagement.

But then you would go a level deeper.

You would ask: what does the human actually do inside each of those areas?

What are all the jobs that exist within the company? What are they doing? How much are they paid?

And to go a layer deeper:

Where do they read things?
Where do they listen to things?
Where do they decide what something means?
Where do they move information from one place to another?
Where do they act as the operating glue holding the whole system together?

That is the work map.

And once you have mapped the work, the next question is: what wedge do you use to attack it?

This is where I think David’s framing is especially helpful, because we are not seeing one single AI wedge work across every market. We are seeing a few recurring patterns.

The first is AI services.

This is maybe the most practical wedge because it goes straight at labor. In a lot of vertical industries, the software exists, but humans are still doing all the real work around it. They are checking, correcting, reconciling, interpreting, and pushing the workflow forward manually. So instead of starting with a better interface, AI services starts with a better way to do the work itself.

Camber is a great example of this. Healthcare reimbursement has historically required huge amounts of specialized human effort layered on top of bad systems and fragmented processes. Claims have to be submitted, fixed, chased, and reworked. Entire services businesses grew up around that pain. What Camber does is turn that labor-heavy reimbursement problem into a software problem. That is why the wedge is so strong. They are not just selling a nicer billing tool. They are going after work that used to require armies of people.

And that is a useful lens for rethinking something like the Toast map. In restaurants, a surprising amount of operational drag still looks like services work hiding inside the business. Catering intake. Guest recovery. AP reconciliation. payroll exception handling. Supplier coordination. These are all places where people are still acting like an internal services layer because the software never really absorbed the work.

The second wedge — and I think the most exciting — is the industry-specific GPT.

This is where things start to get really interesting.

Because now you are no longer just automating a step in the workflow.
You are building an intelligence layer for the vertical itself.

That is a much bigger idea.

An industry-specific GPT is not just a chatbot with a little fine-tuning. It is not just generic AI wearing the clothes of a vertical. At its best, it starts to understand the actual language, logic, edge cases, constraints, and judgment patterns of the domain. It behaves less like a feature and more like an emerging operating intelligence for the industry.

OpenEvidence is a great example of this. The reason it is powerful is not that it is “ChatGPT for doctors.” That framing is too shallow. What makes it compelling is that it is becoming a medical intelligence layer grounded in the realities of clinical decision-making. It is useful because it understands the domain well enough to help practitioners reason inside it.

And once you start thinking this way, the workflow map changes dramatically.

Because now the question is not just: where can I automate labor?

The question becomes: where does this industry need a brain?

What would the restaurant equivalent of that look like?

It would not just answer questions about menu items or write a promo email.

It would understand the operating reality of restaurants. It would understand labor pressure, table turns, reservation pacing, kitchen bottlenecks, menu mix, guest recovery, promo strategy, inventory risk, staffing tradeoffs, refund behavior, and the way all of those things interact.

That is a completely different category of software.

At that point, you are not just building around the workflow. You are building a domain-native intelligence layer that can increasingly guide the workflow.

That is why I think this wedge is the most interesting one.

If the AI services wedge helps you do the work, the industry-specific GPT wedge helps you understand the work at a much deeper level than old software ever could.

And in the long run, that may be the more powerful position.

The third wedge is building agents on top of the system of record.

This one is a little more tactical, but it is incredibly important because it acknowledges a simple truth: in many verticals, the incumbent system still has the data gravity.

It may be hated. It may be clunky. But it is embedded. It is where the records live. It is where the workflow officially lands. That means one of the smartest ways to wedge in is not to replace it immediately, but to start from it.

Use the system of record as the foothold. Read from it. Write back to it. Make it smarter. Reduce the labor around it. Turn it from a passive database into something that feels more action-oriented. You essentially build Agents on top of it.

Datagrid is a great example of this pattern. It does not begin by asking the construction company to replace the system of record (IE Procore). It starts from the fact that the CMS is already embedded in the workflow and already the source of truth. The wedge is to capture the conversation, structure the documentation, and make the surrounding workflow dramatically more usable while integrating into the existing record system. That is a very different strategy from trying to rip out the incumbent on day one.

There is a very obvious parallel in restaurant software too. If you were attacking the Toast map this way, you might not begin by replacing Toast or the broader stack. You might start by asking what AI can do when it sits on top of the systems that already exist. Can it absorb the labor around them? Can it turn raw operational data into clear actions? Can it surface exceptions instead of forcing operators to dig? Can it make the existing stack feel much more intelligent without forcing a painful rip-and-replace?

That is wedge number three.

And the important thing is that these wedges are not mutually exclusive. In fact, the best companies will probably move through them over time.

They may start with AI services, because that is where the urgency and labor budget live.

Then they may deepen into an industry-specific GPT, because that is where the real intelligence advantage gets built.

Then they may sit on top of — or increasingly absorb — the system of record, because that is where the workflow gravity and defensibility live.

That progression feels right to me.

And it also gives us a much better way to redo the original Toast exercise.

The old version of the exercise asked: what are all the categories make up the restaurant?

The new version should ask: where is the labor, what kind of work is being done, how much does it cost, and what wedge is best suited to attack it?

Is this work best attacked as an AI service, because humans are still doing it manually at scale?

Is it a place where an industry-specific GPT can become the intelligence layer for the vertical?

Or is it best attacked by building an agent on top of the system of record, because the workflow gravity is already there and the wedge is to make it dramatically more useful?

That is the strategic map I would want to draw now.

Not just where the workflows are.

But where the work is.
How the work gets done.
And what kind of AI wedge can most naturally absorb it.

That feels like the real AI-era version of the Toast map.

And I think founders who do this well are not going to build a handful of AI features around old vertical software.

They are going to build the new operating layer for the vertical itself.

Do me a solid and forward to a friend :-)

There’s a specific kind of startup worth paying attention to right now: the one that goes straight at a category everyone else assumes is locked up.

Not glamorous markets. Not greenfield software. Markets with incumbents. Markets with procurement. Markets with ancient systems of record. Markets where customers have been overpaying for mediocre software for years.

That’s the backdrop for this episode’s conversation with Alex Niehenke of Scale Venture Partners. The core idea is simple: vertical AI is having its moment not because incumbents are small, but because many of them are slow, bloated, and structurally unable to reinvent themselves fast enough. When the dominant players have low customer love, rising prices, and little appetite to disrupt their own revenue streams, the opening for startups gets real very quickly…

Watch The Episode

Listen To Episode Here

This Weeks Titan: Alex Niehenke (Partner at Scale)

Alex Niehenke is a Partner at Scale Venture Partners, where he focuses on early investments in vertical markets where incumbents have failed to invest in advanced technology. He has been at Scale since 2012 and has backed companies including Motive, Root Insurance, Scout RFP, Range, Proscia, and Archipelago. His investing lens was shaped early by watching his father’s transportation businesses intersect with the rest of the economy, which gave him a long-running curiosity about how real industries actually work.

That background matters because Alex has spent years leaning into sectors many investors historically dismissed as “boring”: insurance, logistics, construction, procurement, financial services, and other operationally dense industries. In the episode, he makes the case that these are exactly the kinds of markets where startups can win big, because the software is often deeply entrenched but poorly loved. The old venture knock on vertical software was always that TAMs were too small and buyers were too hard. Alex’s view is almost the inverse: those markets often only look unattractive until someone finally builds something customers actually want.

Scale’s own positioning maps cleanly to that worldview. The firm describes itself as backing early-stage AI companies on the journey from founder-led growth to a go-to-market machine. In other words, Scale is not just underwriting raw technical novelty; it is looking for companies that can turn product advantage into durable commercial momentum. That framing shows up all over Alex’s comments in the episode, especially in his emphasis on growth velocity, category timing, and whether a startup can become indispensable before the incumbent fully wakes up.

The most useful way to understand Alex is that he is not simply bullish on AI. He is bullish on what happens when AI meets neglected vertical markets. He is looking for wedges where legacy vendors have spent years extracting price instead of reinvesting in product, where user NPS is poor, and where a startup can build what he calls a “system of love” rather than just another system of record. That distinction is the heart of the episode.

Watch The Episode

Listen To Episode Here

1. Build a “system of love,” not just a better workflow

Alex’s most important point is that the best vertical AI companies do not just replace software. They replace resentment. In many verticals, the legacy leader still owns the system of record, but it no longer owns user trust. Customers are stuck with bloated tools, poor UX, and annual price increases that feel disconnected from product improvement. That creates the opening. The startup wins by building the product users actually want to use — the one that feels faster, smarter, and more aligned with how work gets done. That is what Alex means by a “system of love.” It is not a cute phrase. It is a wedge strategy. If the incumbent has spent years training customers to feel captive, then delight itself becomes disruptive.

The best way to frame this in the piece is that vertical AI is often entering markets with structural dissatisfaction already baked in. Alex points to sectors like legal and insurance, where incumbents such as Westlaw, LexisNexis, and Vertafore have long benefited from limited competition, low NPS, and pricing power. In those markets, the startup does not need to convince buyers that the old experience is broken. Buyers already know. The founder’s job is to convert that frustration into adoption by making the new tool feel indispensable almost immediately.

You could sharpen the blog language here by saying: the first product advantage in vertical AI is emotional, not technical. Users do not switch because they got a slightly better dashboard. They switch because the new product respects their time, reduces pain, and gives them leverage against a vendor they never really loved in the first place. That is how small startups start to destabilize giant categories.

2. Use customer hatred as a GTM advantage

One of the hidden advantages in vertical software is that bad incumbents create their own demand for disruption. Alex’s observation is that in a lot of these categories, founders are not selling into neutral markets. They are selling into markets where the customer already feels overcharged, underserved, and trapped. That means the go-to-market motion can be much more direct. The message is not “here is a new tool.” The message is “here is your exit.”

This is especially powerful in categories where incumbents have used contract lock-in, proprietary data, or integration bottlenecks to preserve share. Alex’s point is that once the buyer believes an alternative is real, dissatisfaction turns into urgency. In legal tech, for example, he notes that newer AI-native entrants are benefiting from years of pent-up frustration with legacy research and workflow vendors. In insurance, the same logic applies where multi-owner, PE-shaped incumbents have often optimized for extraction over product improvement. The emotional state of the customer becomes part of the distribution strategy.

A clean way to write this in the post: customer dissatisfaction compresses the sales cycle. When a founder is entering a market where users already hate the default option, education is less important than credibility. The question is not whether the pain exists. The question is whether the startup is trustworthy enough to become the next default.

A quick word from our incredible partner,
Parafin - the leading embedded capital partner.

• 4.8 star rating on Trustpilot, with over 500 reviews [source]

• 71% CSAT rating across all partners

• 84 NPS

• Serving top Fortune 500 companies, including Amazon, Walmart, and DoorDash

Check Out Parafin

3. Don’t fight incumbents in court. Route around them

Alex is very clear on this: founders should not waste their best years trying to litigate their way through a blocked market. His advice is blunt — get the lawyers out of the room. For startups, legal warfare is usually a trap. Incumbents have more money, more patience, and far more ability to absorb distraction. Even when the startup is morally right, the process can still kill momentum.

The better response is tactical. If an incumbent blocks access, restricts integrations, or tries to use data captivity as a moat, the startup should look for ways to become useful anyway. That can mean workflow-based entry, lightweight implementation, partial automation, partner-led adoption, or what Alex loosely refers to as working around the old system rather than begging permission from it. In other words: do not challenge the castle head-on if you can quietly build the road users actually prefer.

This matters because many of the most attractive vertical AI markets are also integration-heavy markets. Alex’s broader framework in his writing is that faster breakout categories tend to require less integration upfront, while slower but potentially more defensible categories require more of it. That means founders need to be strategic about where they need permission and where they do not. The startup that can deliver real ROI before a heavyweight integration is often the startup that earns the right to go deeper later.

4. Win the wedge before you win the platform

A lot of founders talk like they are replacing the incumbent suite on day one. Alex’s worldview is much more practical. The real winners usually start by owning a narrow, painful, high-frequency workflow — something acute enough that users will adopt quickly, even if the broader system stays in place for a while.

That logic is reinforced by his public framework on vertical AI adoption. The earliest breakouts often happen in categories where value is obvious without deep integration and where the sector has historically underinvested in modern software. Legal is a prime example: immediate ROI, text-heavy work, and relatively low implementation friction compared with deeply entangled back-office systems. By contrast, sectors like insurtech may offer huge opportunity, but adoption is slower because the workflows are intertwined with legacy systems and downstream dependencies.

In the post, I’d phrase this as: great vertical AI companies earn expansion through usefulness. First they solve one painful job better than anyone else. Then they become the preferred interface. Then they accumulate context, trust, and data. Only after that do they start to look like a platform. Founders lose when they pitch total replacement too early; they win when they become impossible to remove.

5. Exploit the incumbent’s structure, not just its product weakness

Alex is not just saying incumbents are slow. He is saying many of them are structurally unable to respond. In the episode, he emphasizes that a meaningful number of vertical software leaders are private-equity-owned, heavily levered, and managed for extraction rather than invention. His phrase is that in many of these businesses, “the ambition has left the room.” That is a devastating condition in an AI transition.

This is why the startup’s real edge is often organizational, not merely technical. Founders can take pricing risk. Founders can ship half-finished but highly valuable products. Founders can cannibalize future revenue in order to win the present. A debt-loaded incumbent often cannot do any of those things. It may know the threat is real and still be unable to answer it with the necessary urgency. That is the kind of mismatch vertical AI founders should be looking for: not categories where the incumbent is stupid, but categories where the incumbent is constrained.

That same logic also explains why some of these markets are newly attractive to venture. If the old winner is financially engineered rather than founder-driven, a modern AI-native company can look small in revenue and still be large in strategic threat. The founder is not just shipping better software. They are attacking an institution whose operating model was optimized for a previous era.

6. Growth matters more than neatness

Alex is unusually candid about the market here: growth is trumping everything. Investors are rewarding speed, acceleration, and signs that a company is escaping into category leadership. That does not mean business quality does not matter. It means that in an AI transition, velocity is being read as evidence of inevitability.

He also talks about “elephant accounts” — very large customer deals that can make young companies look lopsided or concentrated, but also signal that serious buyers are willing to place meaningful bets early. For vertical AI founders, this is important. The market is often less interested in perfectly balanced metrics than in whether the startup is becoming strategically important to major buyers. A weird-looking company with explosive pull can be more investable than a tidy one growing at a merely rational pace.

In blog form, the takeaway is: in frontier categories, momentum is a moat. Once a startup becomes the company everyone in a vertical is hearing about, recruiting improves, fundraising improves, customer urgency improves, and incumbent anxiety increases. In these markets, speed does not just create growth. It compounds strategic leverage.

7. Commoditization is coming, so own the workflow, not the feature

Alex is also clear-eyed about the other side of the story. AI lowers the cost of building. What looks novel this quarter may look generic next quarter. He warns that if your company is basically a thin layer on top of model access, pricing pressure will come fast. The consequence is that vertical AI founders need to build defensibility somewhere other than surface-level novelty.

His broader writing offers a useful extension of this idea. In “LLMs and Sneaky Big Markets,” Alex argues that AI changes the economics of vertical software because tools are no longer just helping workers do their jobs — they are beginning to automate pieces of the work itself. That expands TAM, but it also raises the standard for differentiation. The winners will not simply provide software seats; they will capture meaningful economic value by automating real labor, owning critical workflows, and embedding into domain-specific processes.

So the right way to write this playbook is: features commoditize, embeddedness does not. The moat is not that your demo is impressive. The moat is that your product sits in the flow of work, touches proprietary data, understands the regulation, and improves with every customer interaction. That is how vertical AI turns from a fast product into a durable company.

8. The biggest markets are often the ones that looked too small before AI

One of the subtler but more important threads in Alex’s thinking is that AI changes what counts as a venture-scale category. Traditional vertical SaaS math often made certain markets look capped because pricing was seat-based and buyer willingness was constrained. But once software begins automating the work itself, the economic ceiling rises. The market is no longer bounded by software budget alone; it starts to pull against labor budget and workflow value.

That matters for the blog because it reframes why vertical AI is so exciting. It is not just that neglected markets are finally getting better software. It is that categories long dismissed as “too small” can become very large when a product directly absorbs meaningful work. Alex’s point is that founders should stop thinking only in legacy SaaS TAM logic and start asking how much human effort, time, and coordination the product can actually replace.

If you’re building in vertical AI, this episode does not stay at the level of slogans. Alex gets specific about where incumbents are weak, why customer dissatisfaction is a real wedge, how PE-owned software companies lose their appetite for reinvention, and what actually makes a vertical AI company defensible when everyone is worried about commoditization. More importantly, the conversation gives founders a sharper way to think about competition: you do not have to outspend the giants, you have to understand where they stopped serving the customer and move faster than they can respond.

It is also interesting because Alex is not talking about vertical AI as a vague trend. He is talking from the vantage point of someone who has spent years investing in sectors like logistics, insurance, fintech, and construction, and who has backed companies such as Motive, Root Insurance, and Scout RFP through Scale Venture Partners. That makes the discussion feel grounded in pattern recognition rather than hype.

So if you’re a founder, operator, or investor trying to understand where vertical AI is actually going — and how startups can win against entrenched legacy players — go watch the episode. There are a lot of conversations right now about AI tools. This one is about strategy. And that’s what makes it useful.

Watch The Episode

Listen To Episode Here

Thanks for reading LINEAR. I reply to every email…
Have any questions, feedback, comments? Let us know, we work for you!

Learn how white-labeled payments, streamlined setup, and flat-rate pricing helped the platform improve the experience while unlocking new recurring revenue. Read the full case study.

Alright, let’s get to it…

A question I got recently was…

How much time/effort would you say your onboarding takes for your vSaaS?
What about normal support?
Does that vary by plan/price tier and approximate ACV?

I love this question because most founders obsess over CAC and ACV on the way in…

But they don’t spend enough time thinking about the delivery burden after the contract is signed.

And in vertical SaaS, that’s where a lot of the bodies are buried.

Here’s the benchmark they sent me:

• ACV: ~$5K

• Sales process: 1–3 hours on Zoom calls to close

• Onboarding: 5–10 hours of training + 5–10 hours of data migration

• Total onboarding: 10–20 hours per logo

• Ongoing support: ~1 hour per month

• Observation: for every 100% increase in ACV, onboarding time only goes up about 20%

Honestly? That’s pretty healthy.

And it highlights something a lot of founders miss:

Implementation should not scale linearly with ACV.

If your ACV doubles and your onboarding doubles too, you’re not building leverage.
You’re just selling a nicer version of services.

The goal is not to eliminate implementation. The goal is to make sure that as ACV goes up, implementation effort grows much more slowly than revenue.

That is exactly what this persons data suggested.

At roughly $5K ACV, spending 10–20 total hours to get a customer live can make sense if:

• the customer retains well,

• support stays contained,

• and onboarding creates a durable moat rather than a recurring headache.

What would worry me?

If the same customer also needed endless custom setup, weekly handholding, or bespoke migration work that never got productized.

That’s where people fool themselves.
They call it “white glove.”
What it really means is “we haven’t built the product yet.”

So how do I think about it?

ACV doesn’t just shape GTM.
It also shapes implementation architecture.

Low ACV often pushes one toward PLG or marketing-led motions; mid-range ACV leans inside sales; high ACV usually forces higher-touch selling. This isn’t law it’s just what I typically see. The same basic logic applies after the sale. Higher ACV generally supports more human touch — but if you’re doing it right, the effort should rise sublinearly, not one-for-one with revenue.

Here’s the simple mental model I’d use:

Tier 1: Low ACV / light-touch

Think small accounts, simpler workflows, lower switching cost.

• Standardized onboarding

• Template migration

• Group training / office hours

• Async support first

• Product does most of the work

If these customers need lots of calls, custom data cleanup, and one-off workflows, your margins are going to get cooked.

Tier 2: Mid ACV / guided onboarding

This is roughly where your example sits.

• A few sales calls are fine

• 10–20 onboarding hours can work

• Some migration help is fine

• Monthly support can stay lean if the product is clear

• The key is turning repeated onboarding work into product, docs, automations, and checklists

This is often the sweet spot in vertical SaaS.

The customer is paying enough to justify real help.
But not so much that you need a mini-consulting firm attached to every new logo.

Tier 3: Higher ACV / high-stakes onboarding

Once ACV climbs, the implementation burden usually shifts from “training users” to “changing operations.”

Now you’re talking about:

• integrations,

• data mapping,

• policy design,

• stakeholder alignment,

• executive reporting,

• and internal change management.

That’s not just onboarding anymore. That’s operational transformation.

And yes, higher ACV accounts absolutely deserve more touch.

But the best vertical SaaS companies still keep control of the process by making implementation modular, repeatable, and tiered.

That’s the magic.

Not “no humans.”
Precise humans, in the right places.

There’s some good external support for this idea too.

One good onboarding playbook I reviewed recommends using ACV thresholds as routing logic: accounts above roughly $5K estimated ACV can qualify for more human-guided onboarding, while larger accounts may justify even more strategic support. It also suggests very different CS coverage models by segment — roughly 1:20 for high-touch enterprise, 1:100 for hybrid accounts, and automation-first for true self-serve cohorts.

That same playbook had one nugget I really liked:

When accounts stall, speed to human help matters a lot. Customers who got help within 4 hours had a 3–4x higher recovery probability than those that waited 72 hours. That tracks with what most of us have seen in practice. Not every customer needs high-touch onboarding — but the right customer, at the right stuck moment, absolutely does.

So if I were answering the original question directly, I’d say this:

My take

For a vSaaS business at around $5K ACV, 10–20 hours of onboarding per new logo plus ~1 hour/month of ongoing support is very reasonable — if retention is strong and the work is becoming more repeatable over time.

And yes, this should vary by plan and ACV.

But the variation should look more like this:

• Higher ACV = more structured onboarding

• Not necessarily dramatically more onboarding hours

• Higher ACV = more strategic support moments

• Not necessarily infinite reactive support

In other words:

Revenue should scale faster than service load.

That’s the game.

Because once implementation effort starts rising faster than ACV, the software model starts leaking.

And when founders tell me they have a $5K ACV product that takes 40–60 hours to onboard every customer, I usually have the same reaction:

That’s not a SaaS margin profile.
That’s a services margin profile wearing a SaaS costume.

Quick PSA:

Hadrius & Why More Vertical Start-Ups Need To Focus On Solving Compliance

Compliance is one of those categories that most founders avoid because it looks painfully boring from the outside.

But if you’ve been around vertical software long enough, you know that “boring” is often where the monster businesses live.

Hadrius is a great example of that.

They’re building AI-native compliance infrastructure for financial institutions — RIAs, broker-dealers, private funds, and compliance consultants. The wedge is straightforward: automate the ugly stuff that nobody wants to do manually anymore — communications review, marketing review, archiving, employee oversight, firm oversight, and trade surveillance — and pull it all into one system of record.

That market is not small.

YC’s company page says 30,000+ financial firms spend a combined $16B+ a year on ongoing SEC compliance, while still relying on consultants, law firms, and legacy software glued together with spreadsheets and manual reviews. That’s the kind of market that looks sleepy until you realize how much pain is buried under the surface.

And the founding story is exactly the kind of thing I love in vertical SaaS.

The team behind Hadrius didn’t just “discover” compliance from a market map. They ran an SEC-registered robo-advisor themselves, Quantbase, and felt the pain firsthand. Out of that pain they built Hadrius in 2023. They later raised a $2M seed round with participation from Y Combinator, Lynett Capital, Singularity Capital, Dorm Room Fund, Unpopular Ventures, and a handful of fintech founders/operators.

This is textbook vSaaS behavior:

You start with a painful workflow.
You solve the thing that is both mandatory and miserable.
Then you expand into adjacent workflows that already touch the same data, the same users, and the same compliance clock.

That’s exactly what Hadrius appears to be doing.

Their product suite now spans Marketing Oversight, Communications Oversight, Employee Oversight, Firm Oversight, and Transaction Oversight. In other words: they didn’t stop at one point solution. They’re trying to become the compliance OS.

And there are some real proof points showing up.

On its current site, Hadrius says it serves 500+ client institutions, covers $5T+ in AUM across clients, saves 19+ hours per user per week, reduces false positives by 90%+, and claims a 10x reduction in marketing review cycle times. In a regulated vertical, those are not vanity claims. Those are “I can justify this purchase to my boss and to an auditor” claims.

The other thing I like here?

They’re not pitching “AI magic.” They’re pitching regulator-ready evidence, zero-data-retention AI, WORM-compliant storage, SOC 2 Type II compliance, RBAC, SSO, and audit-readiness. That matters. In vertical software, especially in regulated markets, the winner usually isn’t the company with the flashiest demo. It’s the one the buyer trusts enough to put in the workflow.

This is the broader lesson.

AI in vertical SaaS gets way more interesting when it stops being a chatbot and starts becoming a compliance operator, a claims operator, a dispatch operator, a billing operator, a credentialing operator.

That’s where the value piles up.
That’s where the data moat gets deeper.
That’s where the expansion revenue starts to show up.

Hadrius is interesting not just because it’s an AI company.

It’s interesting because it’s attacking a vertical where the pain is mandatory, the workflow is recurring, the switching costs can get real, and the trust layer matters a lot.

And if they keep moving from wedge to workflow suite, you can start to see the shape of a very serious business.

Early Steps In The Transformation To An AI Company

A lot of software companies are still treating AI like a feature.

Add a copilot.
Add a summarizer.
Add a “generate with AI” button somewhere in the UI.

Cool. Fine. Better than nothing.

But that is not the same thing as becoming AI-native. And I think this is where a lot of incumbents are going to get smoked.

The companies that win this next wave won’t just sprinkle AI on top of old workflows. They’ll rebuild the workflow itself.

That’s the big idea behind Intercom’s transformation, which Eoghan McCabe and the team have talked about pretty openly. Intercom moved from legacy SaaS assumptions toward an AI-first operating model by reorganizing teams, re-architecting parts of the codebase, building rigorous internal eval systems, and shifting pricing toward outcomes.

So if you’re running a vertical SaaS company today and you’re not AI-first yet, here’s the playbook I’d follow.

Step 1: Pick one workflow where labor is still doing the work

Not one feature. One workflow.
That’s the key.

Find the place in your product where your customer is still paying with labor:

• reviewing claims

• checking compliance

• writing notes

• calling patients

• routing jobs

• reconciling invoices

• answering repetitive messages

• moving data from one system to another

That is your wedge.

Hadrius didn’t start with “let’s add AI to finance.”
They attacked recurring compliance work that was already mandatory and already painful. That’s why the wedge makes sense.

Step 2: Create a small AI strike team with a single owner

One of the smartest lessons from Intercom is organizational, not technical.

They centralized AI talent, created focused workstreams, and assigned a Directly Responsible Individual to each mission-critical effort. No mushy committee ownership. No “everyone owns it.”

If you want to become AI-native, do not bury AI inside your normal roadmap grooming process.

Spin up a team.
Give it a workflow.
Give it a number.
Give it a boss.

Step 3: Rebuild the system so AI can actually operate

This is where most companies quit.

They want AI-native output from a legacy stack that has:

• messy data,

• brittle workflows,

• poor instrumentation,

• weird permissions,

• and no clean system of record.

Intercom’s team has talked about re-architecting core systems and even modernizing parts of the stack because AI leverage depends on the underlying foundation. Their point was simple: AI is not an add-on. The foundation matters.

In vertical SaaS, this usually means:

• cleaning workflow data,

• tightening permissions,

• exposing the right actions through APIs,

• building evaluation infrastructure,

• and defining what “good” actually looks like.

No clean data, no real AI operator.

Step 4: Prototype internally first, then with design partners

Another Intercom lesson I like: prototype early, test internally, then go out to market with real conviction.

They built internal eval systems, tested against historical data, and ran large-scale experiments before opening things up more broadly. They also used internal teams as early users before putting unstable systems in front of customers.

This is how I’d do it in vertical SaaS:

• Phase 1: internal dogfooding on your own workflows

• Phase 2: 3–5 design partners

• Phase 3: narrow commercial rollout

• Phase 4: broader launch with clear success metrics

Do not skip the eval layer.

If you can’t measure resolution quality, speed, accuracy, hallucinations, or downstream impact, you don’t have an AI product yet.

You have a demo.

Step 5: Price the outcome, not the seat

This is one of the most important mindset shifts.

Intercom’s AI product, Fin, pushed outcome-based pricing — charging for a successful resolution rather than a seat. If the AI escalates and doesn’t finish the job, the customer doesn’t pay for the resolution. That changes the internal culture. Suddenly everyone is aligned around actual customer value. This is a lot easier to say than do, but give it a shot…

In vertical SaaS, that idea gets very interesting.

What if you charged for:

• claims resolved,

• compliance reviews completed,

• invoices reconciled,

• messages handled,

• appointments booked,

• or dispatches completed?

Not every product can move all the way there.
But every founder should at least ask the question.

Because AI-native businesses are often much closer to outcome economics than legacy SaaS businesses.

Step 6: Change the org, not just the product

This is the part people don’t want to hear.

Becoming AI-native usually means changing:

• hiring,

• team structure,

• success metrics,

• product rituals,

• and what “good” looks like across engineering and GTM.

Intercom described this as a real operating-model shift, not a feature launch. They centralized talent, emphasized speed and experimentation, and hired for versatility while still building deep specialist capability where it mattered. BVP Atlas

That’s the opportunity for vertical SaaS founders.

You already have the domain context.
You already understand the workflow.
You already know where the labor is hiding.

Now you need to rebuild your company around eliminating that labor in a trustworthy way.

That’s the difference between “AI-enabled” and “AI-native.”

One gets you a press release.
The other gets you a new company.

And if you’re feeling behind on this, don’t panic.

The good news is vertical SaaS founders usually have a few huge advantages over horizontal teams:

We know the workflow better.
We know where the bodies are buried.
We know what “good enough” actually means in the real world.
We know where the human still has to stay in the loop.

That matters a lot.

So if you’re not AI-native yet, start with one painful workflow, one great team, one measurable outcome, and one real design partner set.

Do that well and you’ll be way ahead of the founders still shipping AI glitter.

You’ve got this.

Do me a solid and forward to a friend :-)

There’s a certain kind of vertical SaaS company that looks small from the outside and massive once you get inside the workflow. This weeks focus company is one of those.

If you looked at it lazily, you’d probably call it booking software for racquet clubs.

That’s technically true, but it completely misses the point. It’s built is much closer to operational infrastructure for an entire category than simple scheduling software.

Let’s dive in…

Listen To Episode Here

This Weeks Titan: Andres Robelo (Founder & CEO at PlayByPoint)

What makes this story interesting to me is not just the market. It’s the shape of the problem.

When Andres started the business, a huge chunk of the racquet sports world was still running on pen and paper, disconnected systems, and generic tools that didn’t really understand how clubs worked. And that’s usually where great vertical SaaS companies are born. They start in markets that outsiders assume are “simple,” but operators know are full of ugly, painful, edge-case-heavy workflows.

That’s exactly what happened here.

A court booking sounds simple until you actually follow the workflow. One player books. Two more get added. Somebody pays later. Somebody else cancels. Refund logic changes based on club policy. Booking windows vary by location. Residential clubs operate differently than country clubs. Municipal facilities have different rule sets than private operators. What sounds like a clean little scheduling problem turns into a pretty deep operating system once you zoom in.

That’s the wedge.

Not “we built booking software.”

More like: we built around a messy, nuanced, rules-heavy operational layer that generic software never cared enough to get right.

And that’s why I think Playbypoint is such a good example of how real vertical titans get built. They don’t win by telling the prettiest story. They win by getting absurdly close to the actual workflow and solving the stuff that looks too annoying for everyone else.

The other thing I really like here is the customer alignment.

Andres is very clear that the company is built around the club, not the end consumer. Yes, there’s a B2B2C dynamic. Yes, there are millions of player touchpoints. But the strategic center of gravity is still the operator. And that matters. Because in vertical SaaS, the operator is the one trusting you with revenue, scheduling, customer data, payments, and the day-to-day functioning of the business.

That trust is what earns you the right to become the system of record.

And once you become the system of record, a whole lot opens up. Payments. Messaging. Retention. Analytics. Marketing automation. AI agents. The expansion paths get a lot more interesting when you own the workflow that actually matters.

One thing that really stood out to me was Andres saying that the business today is still primarily SaaS, not just payments. I expected the mix to skew much more heavily toward payments, but it hasn’t. That tells me there’s real discipline in the model. A lot of founders would be tempted to over-rotate into payments because it looks like the easiest monetization lever. But building a business where your whole moat is take rate compression is dangerous. If somebody better capitalized comes in and buys down the economics, what’s left?

That’s why this company feels thoughtful to me.

It’s not just monetization plumbing.

It’s software depth, operator trust, and category knowledge.

And then there’s the AI angle, which I think is where this story gets even more relevant.

Andres said something that I think a lot of vertical SaaS founders know deep down but don’t always want to say out loud: the old “it took us five years to build this” moat just doesn’t hit the same anymore. That doesn’t mean product knowledge doesn’t matter. It matters a lot. But when building software gets easier, the durable advantage shifts away from code volume and toward workflow understanding, trust, speed, and judgment.

In other words, the question is no longer just “can someone build this?”

The question is “will they know what to build, in what order, with what customer trust, and with what distribution?”

That’s a much better question.

And honestly, it’s why I’m still very bullish on vertical SaaS.

Not lazy vertical SaaS.

Not surface-level vertical SaaS.

But the kind that lives so close to the customer that it understands the business better than anyone else in software possibly could.

Playbypoint feels like that kind of company to me.

It also fits a pattern I’ve seen over and over again in vertical software: the market always looks too small until it doesn’t. Racquet sports, padel, club software — all of that sounds niche until you realize you’re sitting inside a growing category with passionate users, complex operations, embedded payments, physical infrastructure, and plenty of room to layer new products on top.

That’s usually where the best opportunities are.

Not in the categories everyone already agrees are huge.

But in the ones that feel narrow until the workflow and expansion vectors reveal themselves.

That’s what this one feels like.

And I think there are a lot more like it.

Listen To Episode Here

If I were pulling lessons from this story and applying them to vertical SaaS more broadly, the first one would be simple: go where the workflow is ugly.

Too many founders still want markets where the product demo looks clean. That’s usually the wrong instinct. The best vertical opportunities often live inside businesses where the workflow is full of exceptions, weird rules, messy handoffs, and operator pain that doesn’t show up on a clean website screenshot. The uglier the workflow, the more likely it is that generic software has failed.

That’s where you want to be.

The second playbook is to stay obsessively clear about who the customer actually is.

This sounds obvious, but a lot of B2B2C companies lose the plot here. They start by serving the operator, then get distracted by the prettier consumer narrative, and eventually end up building a product strategy that doesn’t serve either side particularly well. The sharper move is to decide who writes the check, who feels the pain, and who decides if you stay or go. In most vertical SaaS markets, the answer is still the operator.

If you get that right, everything gets cleaner. Product. Positioning. Sales. Expansion. Even AI roadmap.

The third playbook is that AI is not killing vertical SaaS. It’s raising the bar.

That’s a very different thing.

If your only moat was that software used to be hard to build, then yes, you should be nervous. But if your moat is real customer intimacy, trust, workflow understanding, and speed of execution, AI can actually widen the gap between strong companies and weak ones. It exposes who really knows the market and who was just benefitting from slower product cycles.

That’s why I think the winners in this next era won’t just be the companies with AI features slapped onto their nav bar. They’ll be the companies that use AI to rethink the actual operation of the customer they serve.

And that leads to the next playbook: build your agent roadmap from the customer’s org chart.

This is one of the cleanest ways to think about AI in vertical SaaS. Don’t start by asking what flashy AI feature you can launch next quarter. Start by asking who works inside the business, what they do all day, what tasks are repetitive, what breaks all the time, and where labor is being wasted on work that should be automated or augmented.

That’s where the real product roadmap lives.

A quick word from our incredible partner,
Parafin - the leading embedded capital partner.

• 4.8 star rating on Trustpilot, with over 500 reviews [source]

• 71% CSAT rating across all partners

• 84 NPS

• Serving top Fortune 500 companies, including Amazon, Walmart, and DoorDash

  Check Out Parafin

  ---

In a business like a racquet club, you can already see it. Failed payment recovery. Retention campaigns. Match recommendations. Reactivation flows. Member outreach. Operator alerts. Staffing coordination. There are dozens of little workflows that used to feel too expensive or too custom to build. Now they don’t.

That changes the game.

Another playbook here is self-cannibalization.

I think this is one of the hardest muscles for software companies to build. When the market shifts, the natural instinct is to protect the old pricing model, the old packaging, the old way value gets sold. But the companies that survive platform transitions are usually the ones willing to break their own model before the market forces them to.

That mindset really matters right now.

If you know your customers are going to care more about outcomes, automation, and proactive workflows than they care about seats and dashboards, then you need to move there before someone else does. You can’t cling to a shrinking definition of value just because it was once the right one.

Only the paranoid survive.

And finally, maybe the most timeless playbook of all: stay embarrassingly close to the customer.

Andres talked about keeping direct WhatsApp groups with customers, and I loved that. It may not be infinitely scalable, but that’s not the point. The point is proximity. Founders lose touch long before they realize they’ve lost touch. The moment you stop hearing the rough edges in real time, you start building for internal narratives instead of real customer pain.

I’ve seen this happen up close.

It’s subtle at first. You still think you know the customer. You still reference old conversations. You still talk like you’re close to the workflow. But little by little, the sharpness disappears. The intuition dulls. And eventually you’re making roadmap decisions from a conference room instead of from the front lines.

That’s when the rot starts.

The best vertical SaaS founders I know all have some version of unreasonable customer proximity. They stay in the pain longer. They keep the feedback loop tighter. They learn faster because they’re still willing to be uncomfortable.

That never goes out of style.

So if I had to boil this week’s playbook down into one line, it would be this:

Find a category outsiders underestimate, earn deep trust with the operator, and move fast enough that your product evolves before the market gets a vote.

That’s the job.

And in this environment, that’s the whole job.

You’ve got this.

Listen To Episode Here

Thanks for reading LINEAR. I reply to every email…
Have any questions, feedback, comments? Let us know, we work for you!

Learn how white-labeled payments, streamlined setup, and flat-rate pricing helped the platform improve the experience while unlocking new recurring revenue. Read the full case study.

Alright, let’s get to it…

InKind: The Most Creative Business Model I’ve Seen In A While…

InKind Is Rewriting The Restaurant Funding Playbook
(And It’s Absolutely Wild)

You know what makes me absolutely giddy about vertical SaaS? When someone takes the most beaten-down, capital-starved, traditionally un-investable industry... and figures out a financial model so creative that it makes VCs who previously ran screaming suddenly start writing checks.

Meet inKind. And buckle up, because this model is chef’s kiss brilliant.

The Founders & The “Why Now” Moment

Johann Moonesinghe and Andrew Harris launched inKind in 2016 in Austin, alongside Johann’s late brother Rajan Moonesinghe and product designer Marcus Triest. But here’s the thing—Johann wasn’t some tech bro trying to “disrupt” restaurants from the outside. The dude owns FOUR restaurants between Austin, Scottsdale, and Las Vegas. He’s lived the pain.

And that pain? It’s restaurant financing. Or more accurately, the complete absence of reasonable restaurant financing.

Think about it: restaurants have historically thin margins, high failure rates, and unpredictable cash flow. Banks don’t want to lend to them. VCs won’t touch them (too asset-heavy, no venture-scale outcome). And traditional equity investors want ownership stakes that most operators aren’t willing to give up.

So what did Johann and Andrew do? They literally cashed out their home and retirement accounts to bootstrap this thing in the early days. Say whaaaaaaaat. That’s conviction, folks.

What It Is & How It Works

Here’s where it gets interesting. InKind isn’t a lender. They’re not taking equity. They’re not even touching your credit card processor.

They’re buying your future food.

Let me break it down: Let’s say you’re a restaurant that needs $100,000 for a kitchen renovation or to smooth out operations. InKind will give you that $100K in cash TODAY—but in exchange, you’re giving them $200,000 worth of food and beverage credit that they can sell to diners through their app.

Then inKind turns around and sells that $200K in credit to consumers through their mobile app—often around $150K, pocketing the spread as their margin. Diners get access to these credits PLUS a 20% cashback reward to use at any other inKind restaurant. So consumers are getting 20 cents back on every dollar spent. Pretty compelling value prop.

For the restaurant, the math works because food cost is typically only 20-30% of menu price. So even selling credits at 50 cents on the dollar, the restaurant can still be profitable—assuming they manage their other costs well and drive enough non-inKind traffic.

The Product Suite: Way More Than Just Financing

But wait, there’s more. (I promise this isn’t an infomercial, though it kinda feels like one.)

InKind has evolved into a full restaurant commerce platform:

1. Sherlock AI - This is their secret sauce. Named perfectly, this system analyzes a restaurant’s menu, local spending behavior, and customer trends to determine EXACTLY how much credit a restaurant can absorb without tanking their cash flow. In Johann’s words from year one: “I lost 50% of the money I funded because I bought $100,000 in donut credits from some donut place in Michigan. It was impossible to sell.” Ouch. But they learned, and now Sherlock prevents that.

2. Labor Management Software - Johann built this internally for his own restaurants and saw a 10% margin improvement. Now they’re rolling it out across their 6,000+ restaurant network.

3. Margin Monitoring Tools - Helping operators actually understand their unit economics in real-time.

4. POS Integrations - They’ve partnered with Toast and other major players so payment processing is seamless.

5. InKind Pass - A monthly subscription for power diners that unlocks premium perks.

The Numbers: Holy Schnikes

Here’s where I had to do a double-take:

• $600M+ deployed to 6,000+ restaurants nationwide

• $450M just raised in February 2026 (Series B led by Magnetar Capital)

• Investors include: Jay-Z’s MarcyPen Capital, ex-Yahoo CEO Jerry Yang, all four members of Metallica (!), and dozens of restaurant owners

• 2% failure rate among funded restaurants (compared to industry average of 60% in first 3 years)

• 4 million+ app users who’ve received $175M in dining rewards

• Partners include José Andrés Group, MINA Group, 20 Michelin-starred restaurants, 50 James Beard nominees

• Estimated revenue: ~$350M (2025 estimates from CB Insights)

• Founders still own 75%+ of the company despite the massive funding

That last stat is WILD. Most founders are diluted to single digits by Series B. But because Johann and Andrew took a relationship-driven funding approach and didn’t need to spray-and-pray with VCs, they maintained control.

Business Model: The Spread Is The Whole Game

InKind makes money on the spread between:

• What they pay restaurants for credit (say, $100K for $200K in credit = 50 cents per dollar)

• What they sell that credit to consumers for (say, $150K = 75 cents per dollar)

• Their gross margin = 25 cents on every dollar of credit purchased

The risk? If restaurants close before the credits are redeemed, inKind eats the loss. That’s why Sherlock AI is so critical—it’s essentially their underwriting engine.

But with a 2% failure rate vs. the industry’s 60%, they’ve clearly figured something out. Part of it is selection (they only fund restaurants with solid fundamentals). Part of it is the inKind effect itself—by driving more customers through the door with that 20% cashback, they’re actually improving the restaurant’s odds of survival.

Capital Raised vs. Capital Deployed

This is fascinating. They’ve raised roughly $600M+ in total funding, and they’ve deployed... $600M to restaurants. They’re essentially a capital-as-a-service business. The funding they raise from investors goes almost directly into restaurant financing, and they recoup it by selling dining credits over time.

Early on, Johann said: “Venture investors hated our business because we’re so balance sheet heavy, we require so much money to give the restaurants. And the debt partners didn’t want to lend us, because they’re like, restaurants are the most risky.”

But now? With a decade of proof, a 2% default rate, and Jay-Z writing checks? The fundraising is relationship-driven, and Johann’s turning investors away.

Outlook For The Future

With that fresh $450M, they’re planning to fund 10,000 additional restaurants in the next year. That would bring them to 16,000+ restaurants.

They’re also expanding the tech platform—more AI-driven tools for labor optimization, margin management, and operational efficiency. Johann’s not just financing restaurants; he’s building the operating system for independent restaurant success.

And here’s the vision: inKind Day. They’re doing a year-long celebration of their 10th birthday with exclusive events in NYC, LA, Chicago, SF, Austin, Miami, and Vegas. They’re building not just a financing platform, but a community and lifestyle brand for restaurant operators and food lovers.

Other Industries That Could Benefit From This Model

Okay, this is where my brain goes into overdrive. Because this model—buying future inventory at a discount, selling it to consumers at a markup, everyone wins—could work in SO MANY verticals:

1. Fitness Studios / Boutique Gyms - ClassPass already does something adjacent, but imagine providing capital to SoulCycle-style studios in exchange for class credits. Sell them to consumers at a discount. Studios get cash flow, consumers get deals, you take the spread.

2. Salons & Spas - Same deal. Buy $200K in haircut/massage credits for $100K cash, sell them for $150K. Built-in customer acquisition for the salon.

3. Auto Repair Shops - This is HUGE. Buy $50K in future service credits, sell them to local car owners. Shops get working capital, car owners lock in discounted maintenance.

4. Veterinary Clinics - Pet owners would LOVE to pre-buy discounted vet visits. Clinics get capital for equipment.

5. Home Services (HVAC, Plumbing, Electrical) - Seasonal businesses with lumpy cash flow. Imagine buying $100K in HVAC credits during winter (slow season), selling them to homeowners before summer (peak season).

6. Dental Practices - Dentists need capital for equipment. Patients want cheaper cleanings and procedures. Match made in heaven.

7. Child Care Centers - Always strapped for cash, parents always looking for deals. Buy tuition credits at a discount, resell them.

The pattern? High-frequency, repeat-purchase businesses with predictable inventory costs in traditionally underfinanced industries. That’s the sweet spot.

The key constraints are:

• Margin structure needs to support the discount (ideally <30% COGS)

• Customer frequency needs to be high enough to sell the credits

• The service/product needs to be appealing enough that consumers will prepay

The Bottom Line

InKind is proof that the best vSaaS businesses don’t just sell software—they solve the existential problems of an industry. For restaurants, that problem isn’t marketing or POS systems (plenty of those). It’s access to non-predatory capital.

By reimagining the entire financing model and building a consumer app to support it, Johann and team created something genuinely new. And with 75% founder ownership at $450M+ raised? They’re playing the long game.

This is what I mean when I say vertical SaaS is about depth, not breadth. InKind went ALL IN on restaurants. They learned the unit economics, built underwriting AI, created a consumer marketplace, integrated with POS systems, and became an indispensable partner to 6,000+ operators.

Could this model work in other industries? Absolutely. But it won’t be easy. You need domain expertise, patient capital, a two-sided marketplace, and a willingness to be balance-sheet heavy.

But if you pull it off? You’re not just building a SaaS company. You’re building the financial backbone of an entire industry.

You’ve got this.

We’re hosting a workshop on
AI-Powered GTM.

This won’t be a theory session. This won’t be a ppt deck full of generic AI takes.

Our guest is going to get into the actual tools and build everything out from scratch in 60 minutes.

Our guest, Ryan Vanshur, VP GTM Intelligence & AI at Handle.com,
and he is one of the most advanced AI pilled folks I’ve met…

If you’re thinking about how AI can actually improve execution across your sales org,
this session will be worth your time.

April 9th, 2026 at 2:30 PM EST

Limited to the first 25 founders/operators.

Sign Up For Free Workshop

Or access our last workshop, Building AI Agents That Actually Work here.

The Best Moat for Legacy Vertical SaaS Systems in the AI Age (And I’ve Seen NO ONE do it yet)

The best moat for legacy vertical SaaS in the AI era might be an agent app store...

If you’re the system of record in a vertical, you already have the thing every AI agent startup wants:

The data, the workflow, the permissions, and the distribution. So why are so few legacy vertical SaaS companies leaning all the way in here? The obvious move isn’t to try to build every AI feature yourself.

It’s to let the market build on top of you.

Give developers and startups governed access to your data + workflow layer
as long as they launch inside your agent app store.

That’s a monster moat.
Why?

Because it means:
— You don’t have to compete with every AI agent upstart
— You don’t have to build an agent for every use case
— You get more innovation than your internal roadmap could ever ship
— You become the distribution layer for the whole category
— You keep the customer relationship and system-of-record status
— You can monetize via rev share, payments, and platform fees

That’s the part I think people are underestimating. In AI, the product moat is getting thinner. But workflow control + proprietary context + embedded distribution? That still compounds.

If I’m a legacy vertical SaaS company, especially a big one,

I’m not only asking:

“Which agents should we build?”

I’m asking:

“How do we become the place where all the best agents get built, discovered, bought, and used?”

That changes the game.

Now every startup building in your vertical is no longer just a threat.
They can become supply.

You don’t need to out-innovate the entire market.
You need to orchestrate it.

And if you do this right, you get the upside of AI innovation without carrying all the product risk yourself. That’s a way better business.

The incumbents already have the ingredients:
-customer trust
-workflow ownership
-historical data
-embedded payments potential
-implementation footprint
-distribution into the exact end user

So again, the question is:
Why haven’t more legacy vertical SaaS brands built AI agent app stores yet?

Feels like one of the clearest moats of this era.

I'm asking you Veeva Systems, Procore Technologies, Yardi, Toast, nCino, etc. (!)

What if the most profitable AI businesses aren’t built by chasing moonshots—but by solving the unsexy, painful problems that bleed money every single day?

Christophe Rimann and Camber didn’t build a general-purpose AI platform. They built a vertical machine so specific, so disciplined, and so ruthlessly focused on one problem that it became nearly impossible to compete with….

Watch Full Episode Here

Listen To Episode Here

This Weeks Titan: Christophe Rimann (Founder & CEO at Camber)

Christophe’s path into healthcare wasn’t linear, which is usually a good sign.

He got into crypto back in the 2012–2013 era, when Bitcoin still looked like fake internet money to most people. He went on to build a crypto broker-dealer that hit about a $1 million run rate and operated with a FinCEN license at a time when almost nobody knew what the rules even were. In other words: he got his reps in early building inside a chaotic, infrastructure-heavy market where trust, compliance, and edge cases mattered.

Then he went deeper technically.

After crypto, he went to grad school to do blockchain-related research, then joined McKinsey, where he spent nearly two years doing healthcare consulting, much of it in revenue cycle management for large hospital systems. That matters. Because once you’ve spent enough time in rev cycle, you realize this isn’t some back-office annoyance. It is the bloodstream of the business. You can have great clinicians, strong demand, and solid patient outcomes — but if claims don’t get paid, none of it works.

💡But the real driver here wasn’t just market logic. It was personal.

Christophe talked about being diagnosed with ADHD in college and how access to behavioral healthcare changed his life. He explicitly said he doesn’t think he’d be in a position to run a company today if he hadn’t gotten that care. That’s not branding fluff. That’s founder-market fit with actual emotional depth behind it.

He didn’t back into healthcare because it was a big TAM slide. He got there because he saw firsthand how important access to care is — and then later saw how broken the financial plumbing behind that care really was.

At big hospital systems, a human can chase the last 5–10% of a claim because it might be worth thousands. But in outpatient and behavioral health, the economics fall apart.

If a clinic gets paid 90% of a $150 claim, are they really going to pay a person to chase the last $15? Maybe. But probably not at scale. And when you multiply that across hundreds of thousands of claims, clinics end up leaving tens of thousands — even millions of dollars — on the table.

🎯 Camber was built on the idea that the long tail of healthcare claims is too operationally painful for humans and too economically important to ignore.

SAY WHAAAAAAAAT.

That’s a very good vertical SaaS wedge. It’s painful, frequent, workflow-native, tied directly to revenue, and hard enough that most software companies never want to touch it.

Watch Full Episode Here

Listen To Episode Here

#1. Pick a wedge where the market already exists — then ask if you can actually build the product

One of the smartest things Christophe said was that Camber didn’t really have a product-market-fit problem. Revenue cycle has existed for decades. Everyone knows the pain. Buyers already spend money on it.

The real question wasn’t “will someone buy this?” It was “can we actually build software that performs at or above the level of resilient human labor?”

There are markets where distribution is the hard part. And there are markets where the pain is so obvious that distribution is not the primary constraint. In those markets, the real bottleneck is capability. Can the product actually do the job? Can it survive edge cases? Can it outperform the incumbent process in live production, not just in a sandbox?

#2. Stealth is not cowardice — it can be sheer discipline

Camber raised pre-seed, seed, Series A, and Series B without announcing a single round. No press releases. No chest pounding. Just building.

Because GTM wasn’t the bottleneck early on. The bottleneck was proving they could build a system that actually worked. That’s a contrarian move in a market where founders are often rewarded for narrative before substance.

Some of the best vertical businesses won’t look “hot” on Twitter during the build phase. They’ll look quiet. Weird. Unfashionable. Maybe even too operational.

Then one day they emerge with real customers, real margins, real data, and a much stronger moat than the loud company that spent three years farming engagement.

A quick word from our incredible partner,
Parafin - the leading embedded capital partner.

• 4.8 star rating on Trustpilot, with over 500 reviews [source]

• 71% CSAT rating across all partners

• 84 NPS

• Serving top Fortune 500 companies, including Amazon, Walmart, and DoorDash

  Check Out Parafin

  ---

#3. Don’t fake automation with labor

Christophe said they made an explicit decision not to offshore the work. Why? Because once you use cheap labor as the release valve, you stop feeling the pain strongly enough to automate it properly. The incentive to fix the root problem disappears.

A lot of founders will tell you they’ve built automation when what they’ve really built is a labor buffer wrapped in software. If labor becomes the permanent crutch, your gross margin story is fake and your product never compounds.

No offshoring. No hiding from the pain. Build the machine. And according to Christophe, that decision led them to gross margins that look like standard SaaS — even upper-quartile SaaS — despite operating in a category most people would assume has to look like a services business.

That’s how real moats get built.

Watch Full Episode Here

Listen To Episode Here

#4. Humans in the loop are not a bug — they’re the training data advantage

There’s a lazy narrative in AI right now that the best system has zero humans involved. Christophe described their insurance operations team almost like expert labelers for a foundation model. When a claim fails or a strange edge case appears, humans don’t just patch the issue. They feed the learning back upstream so the system improves.

Your goal is not to eliminate human expertise on day one. Your goal is to convert human expertise into proprietary system intelligence faster than anyone else.

#5. Sell EBITDA uplift, not just cost savings

Most AI companies sell some version of labor reduction: fewer heads, lower costs, faster workflows. Camber goes after yield. They care about collections. They care about getting claims paid at a higher rate. They care about increasing EBITDA, not just reducing payroll.

If you help a provider collect an extra 100 to 300 basis points, that’s not some nice-to-have dashboard improvement. That can be a deeply material financial change in a tight-margin business.

They don’t charge a classic SaaS fee. They charge a base fee tied to collections — usually lower than what the customer is already paying — and then participate in the upside they generate. The customer wins first. Then Camber wins with them.

#6. Don’t become the system of record if the better business is the payments layer

Do you need to own the system of record to matter? Camber’s answer is basically: not necessarily. Christophe’s view is that they’re building something much closer to Stripe for healthcare — a payments network that sits on top of systems of record rather than replacing them.

In a lot of vertical markets, the system of record gets all the attention, but the economic value ends up accruing in the payments layer, financing layer, or workflow infrastructure wrapped around it.

Are you really supposed to own the whole stack? Or is there a narrower layer where you can become indispensable to the entire ecosystem?

#7. Going vertical is what makes AI actually work

Camber started in behavioral health, then expanded into areas like physical therapy, neuropsychology, ENT, and cardio. That expansion didn’t happen by accident. It happened because the model is learned specialty by specialty.

People assume broader is better because general models are broad. But in regulated, workflow-heavy industries, breadth is often a weakness.

#8. Distribution can come from private equity before it comes from the market

Christophe said one of Camber’s most common GTM motions is to talk to private equity firms first. In some cases, PE groups bring them into diligence before buying an asset so Camber can help identify revenue and EBITDA upside.

If your product can materially improve margins, reduce leakage, or tighten operations inside fragmented service businesses, you may not need to start with the operator. You may be able to start with the capital allocator.

A software vendor that can plug directly into diligence, quantify upside, and then help realize it post-close is not just a tool vendor. That starts to look like strategic infrastructure.

#9. In enterprise vertical AI, change management is part of the product

This is such an important reminder for vertical AI founders who think better models alone will win the enterprise. If you are selling outcomes, then you are selling workflow change. And workflow change means training, implementation, trust-building, process redesign, ownership alignment, and a lot of on-the-ground work.

That doesn’t make your business less software-like. It makes it more real. The companies that embrace this will outperform the ones that pretend implementation is someone else’s problem.

Why This Matters for the Rest of Us

The big idea here isn’t just that Camber is automating claims. It’s that they found a painful, fragmented, high-frequency workflow sitting directly next to revenue… and then they built a data engine around it before the rest of the market woke up.

The next generation of great vertical SaaS companies may not look like traditional SaaS at all.

They may look like services on the surface, software underneath, and financial infrastructure at the core. That’s a very interesting place to build — and a very interesting place to invest.

You’ve got this.

Watch Full Episode Here

Listen To Episode Here

Thanks for reading LINEAR. I reply to every email…
Have any questions, feedback, comments? Let us know, we work for you!

Today’s newsletter is sponsored by Xplor Pay. In their latest case study, they show how a POS provider embedded payments using PayFac as a Service without taking on the burden of compliance, risk, or merchant onboarding.

Learn how white-labeled payments, streamlined setup, and flat-rate pricing helped the platform improve the experience while unlocking new recurring revenue. Read the full case study.

Alright, let’s get to it…

Why OpenClaw Might Be a Bigger Vertical SaaS Opportunity Than Most People Realize

Most people look at OpenClaw and see a personal AI assistant.

I think that framing is too small.

When a new AI product breaks out, most people try to put it into one of two buckets. Either it’s a toy, or it’s a wrapper. That’s usually the whole conversation. “Cool demo, but who cares?” Or, “Interesting product, but what’s the moat?” And honestly, that’s often the right instinct. We’ve all seen enough thin AI products to know how this usually goes.

But OpenClaw feels a bit different.

Not because it magically solves trust, workflow design, distribution, or any of the hard stuff that actually makes software valuable. It doesn’t. But because if you look at what it actually does, it doesn’t really feel like a normal chatbot. It feels more like an execution layer.

And once you see that, the vertical SaaS angle gets a lot more interesting.

The wrong way to think about OpenClaw

If you describe OpenClaw as “a personal AI assistant,” you’re not wrong.

You’re just leaving out the most important part.

That description makes it sound like another chatbot with memory. Something that sits in a chat window, answers questions, drafts a few things, maybe remembers some context, and helps you feel a little more productive. That’s fine, but it undersells what makes it potentially valuable.

The interesting thing about OpenClaw is not that it can respond.

It’s that it can operate.

It can sit across tools, channels, and workflows and actually do things. It can move work forward. It can exist closer to where business processes actually start, which is usually not inside some beautifully organized dashboard.

That matters a lot more than people think.

Because businesses do not really pay for “AI that talks nicely.” They pay for outcomes. They pay for fewer dropped balls, faster response times, more booked appointments, cleaner handoffs, lower labor costs, and more revenue collected. If an AI product cannot move one of those numbers, it may still be cool, but it’s probably not a business.

OpenClaw at least points toward that world.

So what is OpenClaw, really?

The simplest way to think about OpenClaw is as an open-source AI assistant that runs on your own devices and works across the tools and channels you already use. That alone is interesting. But the more useful framing is this: OpenClaw behaves less like “a bot you talk to” and more like “a system that can sit inside your operational environment.”

That’s a big distinction.

Most people imagine work happening inside the app. But if you’ve spent time in real businesses, especially messy vertical ones, you know that’s not how it works. The app is often where the record ends up. The actual work starts before that, around that, and between that.

It starts in:

• text threads

• missed calls

• support inboxes

• Slack messages

• WhatsApp conversations

• scheduling requests

• random follow-ups nobody owns

That messy layer is where a lot of the pain lives. It’s also where a lot of software still does a terrible job.

That’s why OpenClaw matters.

It’s not just another AI UI. It looks more like a programmable layer that can live inside the chaos and help convert communication into action.

Why that matters for vertical SaaS

If you’ve built in vertical SaaS, or even just studied the good companies long enough, you start to notice the same pattern over and over again.

The best businesses usually do not start with some giant category-defining vision.

They start with one painfully specific workflow.

One workflow that is:

• frequent

• expensive

• repetitive

• full of manual steps

• tied to money or labor

• annoying enough that customers will gladly pay to make it go away

That’s the whole wedge.

Not “software for dentists.”

Not “AI for real estate.”

Not “digital transformation for SMBs.”

One ugly workflow.

That’s where vertical winners are born.

And OpenClaw gets interesting because it seems well-suited for exactly that kind of wedge. It maps well to environments with fragmented systems, lots of communication, too much manual follow-up, weird handoffs between people and tools, and revenue leaking out because no one owns the operational middle.

That’s basically vertical SaaS in a nutshell.

The key idea: OpenClaw as a workflow layer

The most useful way to frame OpenClaw is this:

It becomes valuable when it acts as the communication and execution layer around a painful workflow.

Not the whole stack.

Not a magical AI employee for everyone.

A focused layer that:

• receives input from the channels people already use

• interprets what needs to happen

• triggers the next step

• routes context into the right system

• keeps the workflow moving until it’s done

That’s a much stronger product concept than “assistant in your messaging app.”

It also fits how a lot of great vertical SaaS products actually get built. They don’t replace everything on day one. They wedge themselves into a high-value workflow, own that workflow, then expand from there.

That’s the opportunity here.

What OpenClaw is not

This is where people get themselves into trouble.

OpenClaw is not automatically a company.

It is not automatically a moat.

And it is definitely not a finished vertical product.

Say whaaaaaaaat.

This is the part people in AI land constantly want to skip. They find a cool runtime or framework and act like the hard part is over. It’s not. The hard part has barely started.

You still have to do the real work:

• pick the right market

• understand the workflow deeply

• build trust

• integrate with incumbent systems

• handle ugly edge cases

• package the UX

• prove ROI

• own distribution

That’s the business.

OpenClaw is the ingredient, not the meal.

Where the use cases get interesting

This is where things start to click.

If you stop thinking of OpenClaw as a generic assistant and start thinking of it as infrastructure for workflow software, some really compelling vertical use cases show up.

1. Missed-call recovery for local-service businesses

There are still a lot of verticals where the phone is the front door to revenue. Dental practices, med spas, HVAC companies, plumbers, lawyers, pest control operators, specialty clinics — these businesses do not treat calls like casual communication. Calls are often purchase intent.

So when they miss a call, they are not just missing a message. They are missing money.

That’s a great wedge.

An OpenClaw-powered vertical product here could answer after-hours inquiries, qualify leads, collect the right details, book appointments, send the follow-up text, and push the context into the CRM or scheduling system. That is a very clean example of where an execution layer creates value.

The buyer does not need to believe in AI.

They just need to believe that missed calls are costing them business.

2. Property management operations

Property management is another obvious one. The workflow volume is constant and the communication burden is enormous. Leasing inquiries, tour scheduling, maintenance triage, resident follow-up, reminders, collections nudges, internal coordination — it never stops.

That makes it a great environment for a conversational execution layer.

The opportunity here is probably not to replace the property management system. That’s usually too ambitious and too expensive at the start. The better opportunity is to own the workflow layer around it.

That means being the system that:

• answers first

• routes correctly

• follows up automatically

• keeps the lead warm

• reduces admin workload

• improves response speed

• helps reduce bad debt

That’s a real product wedge. And if you get it right, you can expand from communication into deeper operational ownership over time.

3. Back-office coordination in ugly industries

Some of the best software businesses are built in the ugliest workflows.

Construction ops. Logistics. Insurance admin. Healthcare back office. Specialty trades. These environments still depend on shared inboxes, spreadsheets, voicemails, status checks, document chasing, and too many humans manually bridging systems that do not talk to each other well.

The problem there is usually not lack of intelligence.

It’s lack of execution.

Someone has to keep the workflow moving. Someone has to follow up, escalate, remind, route, and confirm. And too often, that “someone” is a tired ops person drowning in tabs and email threads.

That’s exactly where an OpenClaw-like layer could shine.

You could imagine products built around:

• document collection

• status chasing

• vendor coordination

• claims follow-up

• dispatch confirmations

• internal escalation workflows

These aren’t sexy demos. They’re better than sexy demos. They’re useful.

4. Payments and collections

This one is sneaky big.

The best vertical SaaS products do not just help customers work. They help customers get paid. And once software gets closer to money movement, the ROI gets way easier to understand.

That’s why payments and collections workflows are such an interesting area for something like OpenClaw. A focused product could handle invoice reminders, failed payment recovery, financing nudges, payment-link follow-up, and escalation flows for overdue accounts.

That is a much easier sale than “AI transformation.”

If the customer sees:

• more recovered revenue

• better cash flow

• less time spent chasing payments

• fewer delinquent accounts

…then the product story becomes very simple.

And simple product stories usually sell better.

5. Field-team copilots

A lot of workers in vertical industries do not want more software. They want less friction.

They want to talk, text, upload a photo, get the next action, and move on.

That’s why field-team copilots are interesting. Think technicians, inspectors, dispatchers, estimators, site managers, care coordinators. In those jobs, the ideal interface is often not “another dashboard.” It’s something much lighter.

An OpenClaw-like layer could help turn informal inputs into structured action. The worker sends a note, a photo, or a message. The system translates that into the next step, routes the context, updates the right system, and keeps things moving.

Less clicking. More doing.

That’s a very strong value proposition in a lot of vertical markets.

What founders should be careful about

This all sounds promising, and I think it is promising. But there are a few places founders can get themselves into trouble.

First, it’s easy to confuse a flexible runtime with a finished product. Customers do not buy “flexibility.” They buy solutions to painful problems. So if you build on OpenClaw, your job is not to show how many things it can do. Your job is to make one important workflow feel inevitable.

Second, trust matters a lot. The second your product touches payments, healthcare, legal workflows, scheduling, or customer communication, “pretty good” is not good enough. Reliability, control, guardrails, and error handling matter. A lot.

Third, the moat is never the wrapper. The moat is the workflow. More specifically, it’s the domain expertise, the system integrations, the edge-case knowledge, the customer trust, the data exhaust, and the distribution. That’s how vertical SaaS companies win. Not by having a clever AI layer, but by becoming deeply embedded in how one industry actually operates.

The bigger opportunity…

The real opportunity here is not to build a horizontal “AI assistant for businesses.”

That market gets crowded fast, and it usually gets commoditized even faster.

The real opportunity is to take OpenClaw-like capabilities and wrap them around a painful, recurring, industry-specific workflow with immediate ROI.

That’s the playbook.

If I were looking at markets through that lens, I’d want workflows that are:

• communication-heavy

• repetitive

• operationally messy

• easy to measure

• tied to revenue, collections, response speed, or labor savings

That’s where this gets really interesting.

Because that’s where software stops being cool and starts being valuable.

To wrap it up…

OpenClaw is easy to underestimate because “personal AI assistant” sounds small. It sounds like a category we already understand.

I don’t think that’s the right way to look at it.

Underneath that label is something more interesting: a system that can sit across communication surfaces, tools, and workflows and help convert messy inputs into action.

On its own, that’s not a company.

But in the hands of a founder who deeply understands one vertical, one painful workflow, and one sharp wedge, it could absolutely become the foundation for a very real business.

Not because it’s magic.

Because it fits the way real industries still work.

And in vertical SaaS, that’s usually where the best opportunities hide.

The Story of Gloss Genius:
The OS For Beauty Businesses

You know what’s wild? There are over 444,000 licensed salons, spas, and barbershops in the US alone. That’s a $247 billion global market. And until recently, most of these businesses were running on... pen and paper.

I’m talking about talented hair stylists, aestheticians, nail techs, and massage therapists who could charge $100+ per hour for their craft, but were stuck managing their entire business with a physical appointment book and Square Reader. Maybe Mindbody if they wanted to drop serious cash on clunky software built for yoga studios.

Enter GlossGenius.

Founded in 2016 by Danielle Cohen-Shohet and (who did makeup artistry as a side hustle at Princeton), GlossGenius is now serving over 70,000 beauty and wellness businesses, reportedly hit close to $100M in revenue (as reported by Axios in 2024), and sits at a $510M valuation. And they did it by going vertical in one of the most fragmented industries in America.

The Wedge: Scheduling + Payments

Here’s what Danielle got right from day one: she understood that time is the unit of currency for beauty professionals. If you’re a hairstylist, your business scales based on how many hours you can book. Every minute spent managing appointment books, chasing down payments, or manually sending reminders is money left on the table.

So GlossGenius started simple — scheduling, payments, light CRM, and basic reporting. All in one elegant mobile-responsive web app. At a time when it wasn’t conventional to integrate payments with the rest of the stack (everyone told Danielle this was a mistake), she bundled it anyway because it was the same damn customer.

The wedge worked. Beta users started emailing support asking if their colleagues could use it too. Word of mouth took off. Beauty pros who normally hated “business stuff” were posting about GlossGenius on Instagram.

The Full Stack Play

But here’s where it gets interesting. GlossGenius didn’t stop at booking and payments.

They kept going deeper:

• Client management with detailed profiles, service history, and notes

• Automated SMS/email for confirmations, reminders, rebooking

• Forms and waivers that auto-populate client files before appointments

• Marketing automation with segmentation by visit history and spend

• Inventory management with barcode scanning

• Payroll that auto-pulls hours, tips, and commissions in three clicks

• Buy Now Pay Later at checkout (6% fee vs. 2.6% standard)

• Genius Loans — pre-qualified capital from $1K to $250K via Stripe Capital

That last one? That’s the vertical SaaS dream. You sit on billions of dollars in transaction data. You know exactly which businesses are healthy. So you can underwrite capital better than any bank.

The Business Model: Subscriptions + Payments

GlossGenius charges $24 to $148/month depending on team size, plus 2.6% on every card transaction. That dual revenue model is beautiful — recurring SaaS revenue plus a growing take rate as customers process more volume.

Sound familiar? It’s the Toast playbook. And just like Toast, about half of GlossGenius’s value comes from the payments side.

The Market Opportunity

Here’s what gets me excited about this space:

1. Massive fragmentation: 29% of beauty professionals are self-employed. Salon suites (where stylists rent private spaces) are exploding. More talented pros are going independent thanks to Instagram and TikTok.

2. Underserved customer: 40% of beauty businesses aren’t using any technology. The ones that are? They’re cobbling together Square + Google Calendar + Mailchimp. Their systems don’t talk to each other.

3. Payments unlock: The industry is finally moving toward upfront payments and hourly billing (vs. per-service). Why should a talented stylist eat the cost of no-shows when you pre-pay for flights and concert tickets?

4. TAM expansion: GlossGenius started with hair but is already expanding into massage ($21.6B), aesthetics, wellness, personal training, yoga. The global wellness economy is projected to hit $9 trillion by 2028.

The Capital Efficiency Story

What I respect most? Danielle kept funding rounds small and stayed capital efficient.

• 2016: Raised a seed from angels with just a few hundred customers and zero revenue

• 2017: Launched paid product, grew 10% month-over-month organically

• 2021: $16.4M Series A from Bessemer

• 2022: $25M Series B (tripled valuation)

• 2023: $28M Series C at $510M valuation

• Total raised: ~$70M to get to an estimated $100M in revenue (as reported by Axios in 2024)

That’s roughly a 1.4x revenue-to-funding ratio if the estimates hold. In a world where vertical SaaS companies are raising $100M+ to get to scale, GlossGenius did it with a fraction of the capital.

How? Product-led growth. Word of mouth. Maniacal focus on customer value. Danielle spent years personally answering support tickets because it gave her direct insight into what customers needed.

The Competitive Moat

GlossGenius competes with:

• Horizontal players (Square, Mindbody) — too generic, missing industry-specific workflows

• Vertical freemium (Fresha) — free software, monetizes via 20% booking fees

• Vertical marketplaces (Vagaro, Booksy) — consumer discovery platforms

GlossGenius differentiated by:

1. Simplicity: Beauty pros love how easy it is vs. clunky Mindbody

2. Pricing: $24/mo beats Boulevard’s $158-$369/mo enterprise pricing

3. Payments integration: Capturing transaction volume competitors miss

4. Trust: They don’t compete with their customers via a marketplace (yet)

What’s Next?

The fintech expansion is obvious. With Genius Loans live, they could add business bank accounts, credit cards, and insurance. They’re sitting on the data to underwrite risk better than anyone.

The marketplace question is interesting. Fresha, Vagaro, and Booksy all drive incremental bookings through consumer discovery. GlossGenius has 70,000+ locations — enough liquidity to launch a marketplace that actually works. But do they want to risk the relationship with customers who might see them as competition?

My bet? They’ll build it eventually. But they’ll be careful about positioning it as a tool for their customers, not a replacement.

The Bottom Line

GlossGenius is a textbook vertical SaaS success story:

✅ Picked a massive, fragmented, underserved market
✅ Wedged in with a simple, must-have product
✅ Went full stack with payments, fintech, and operational tools
✅ Built a product customers love enough to recommend
✅ Stayed capital efficient while scaling to estimated $100M ARR (as reported by Axios in 2024)

And they’re just getting started.

You Can MONOPOLIZE In Vertical SaaS

You CAN monopolize in Vertical Software.
We’ve seen it time and time again: your target should be 100% market share.

Not because you’ll literally get every customer in the market.

But because aiming for anything less usually leads to building a nice tool… instead of a category-defining company. That’s the big misunderstanding in vertical SaaS.

People think,
“this is a niche market”
or
“if we get 5-10% share, we’ve done great.”

Maybe.

But the best vertical software businesses don’t think that way.

They ask:
How do we become the default operating system for this industry?
That means more than just solving a few pain points.

Owning the data.
Owning the payments.
Owning the daily habit.
Using AI to actually SELL the work to be done.

That’s how vertical software compounds.

You start with a wedge: scheduling, CRM, estimating, compliance, charting, dispatch, inventory, whatever the market is begging for.

Then you expand.

More workflows.
More users.
More products.
More revenue per customer.
Higher switching costs.
Deeper moats.

Eventually, you’re not selling software.
You’re running critical infrastructure for that industry.

That’s why the great vertical companies feel so dominant.
Not because the market is massive at the start.

But because they go deep enough to make themselves unavoidable.

The question isn’t:
“Is this TAM too small?”

The better question is:
“If we win this vertical completely, how big does this get?”

That’s usually where things get interesting...

When Deepak Chhugani started Nuvocargo in 2019, the startup HATED tech-enabled services. Then came the SaaS-only correction. Then the agent gold rush. Through every hype cycle, Deepak kept building toward one thesis: customers pay for outcomes, not tooling.

In this deep-dive conversation on the Verticals podcast with host Nic Poulos, Deepak breaks down exactly how Nuvocargo went from a scrappy YC startup moving tequila across the Mexican border to an AI-native logistics platform with $85 million in funding, multiple acquisitions, and a product that’s making traditional freight brokers rethink everything.

Along the way, he delivers a masterclass in why most AI agent startups are failing to retain their customers — and what to do instead.

Watch Full Episode

Listen To Episode

Vertical Titan: Deepak Chhugani

Deepak Chhugani is the founder and CEO of Nuvocargo, an AI-native logistics platform powering North American freight. Born to Indian parents and raised in Ecuador, his multicultural upbringing gave him a rare edge in a market defined by cross-border complexity. He attended school in Latin America before coming to the US in 2010, where he launched his career on Wall Street at Bank of America Merrill Lynch.

In 2015, a twist of fate reshaped his trajectory. Deepak was one of thousands of immigrants who lost the H-1B visa lottery — “a third of folks in these high-paying jobs had to leave the country,” he recalls. Rather than let him go, BofA relocated him to their Mexico City office, where he spent years helping large family businesses and industrial companies raise capital, think about IPOs, and structure deals. That exposure to the Mexican industrial economy, combined with his family’s logistics business back in Ecuador, planted the seed for what would come next.

When he left banking, Deepak entered the startup world through a winding path — a mix of early-stage ventures and private equity, “1 to 2 years figuring out what the hell I’m doing,” as he puts it. He got into Y Combinator, went through pivots and idea changes, and landed on the concept that became Nuvocargo: a technology-driven logistics company focused on the massive, underserved US-Mexico trade corridor.

He was 26 when he founded the company in 2019. Seven years later, Nuvocargo has raised over $85 million from Tiger Global, QED, NFX, YC, and angels including the founders of Ramp, Nubank, and Deel. The company has made multiple acquisitions — in software, AI, freight brokerage, and customs — and expanded from US-Mexico cross-border to all of North America. Revenue is in the mid-tens of millions, and the company was recently named one of Fast Company’s most innovative startups.

What makes Deepak unusual among logistics founders is the combination: a bicultural, bilingual background that gives him authentic credibility on both sides of the border; a finance pedigree that taught him capital structure and deal-making; and a first-time founder’s willingness to challenge conventional wisdom about how freight should be built and sold. “A lot of companies that were doing tech-enabled services in freight had a lot of challenges post-2022,” he says. “I think that we weathered this crazy freight recession really, really well, and we reinvented our business for the age of AI.”

From Wall Street to the Border

The US-Mexico trade corridor is a trillion-dollar market — $50 billion spent annually on freight alone — but unlike most international freight, it all moves over the road. Not air. Not ocean. Trucks. Deepak saw a unique niche where his bilingual, bicultural background gave him an edge no one else had.

“It was a unique niche,” Deepak says. “I thought it would be a great place to start.” He went through Y Combinator, pivoted a few times, and landed on what would become Nuvocargo — a company that would grow from a handful of cross-border tequila shipments into a platform managing freight across all of North America.

“Every little thing around you has parts or was moved in trucks and goes through the global supply chain. Trade is what moves the world. Trade is what connected populations.”

— Deepak Chhugani

The First Shipment: A Tequila Brand and a Stack of Immigration Papers

Nuvocargo’s origin story has the kind of serendipity that can’t be manufactured. Deepak’s very first customer was Carlos Soto, the founder of Nosotros Tequila — a small (at the time) tequila brand importing from Mexico into the US.

How did he win the account? Not through a pitch deck or a demo. Months earlier, when Deepak had no business relationship with Carlos, he’d gifted him a stack of immigration paperwork that had cost him over $10,000 — documents he no longer needed, but that would save Carlos enormous time and money navigating the same immigration process. “I gave him all this paperwork just to be nice because I knew the immigration process was very annoying,” Deepak explains.

When Nuvocargo launched and Deepak reached out about handling some cross-border shipments, Carlos remembered. He was “extremely helpful” and gave Deepak a shot with a few early loads. From there, it was network effects and “good old-fashioned hustle” — customers in plastics, packaging, manufacturing. Names most people wouldn’t recognize, but shipments that taught the team how freight actually works at the ground level.

“They started trusting us with a few shipments here and there,” Deepak says. “And then we started learning the process and really building software and tools to make that whole experience better and better.” The lesson embedded in this origin: in B2B services, generosity and trust compound over time in ways that cold outreach never can.

The Bilingual, Bicultural DNA

The US-Mexico freight market has a peculiar structure that shapes everything about how Nuvocargo operates. Unlike a typical domestic freight play, decision-makers sit on both sides of the border. A Mexican manufacturer exporting to the US might be responsible for delivering to an American warehouse. An LA-based brand importing tequila needs someone to handle transportation from Mexico. You’re selling to both types of customers simultaneously.

“That’s why the company’s DNA was very bilingual, bicultural from day one,” Deepak explains. “And that was really one of our unique advantages before things around our software or our regulatory compliance and licenses became more of our powerful differentiators.”

When asked what advice he’d give founders building a cross-border business from day one, Deepak is candid about the difficulties. He was 26 and a first-time founder. COVID hit in their second year, shattering any plans for in-person culture building. “Despite that, we were pretty good about keeping two main hubs, New York and Mexico City. I would say it’s hard.”

But the cross-border structure was also a survival mechanism. Companies that built all their operations in expensive US cities like San Francisco, Seattle, or New York ended up with “incredibly bloated cost structures.” Nuvocargo’s distributed model — born out of necessity — turned out to be a structural advantage. “If you had to ask me, like, my wave a magic wand, how would I have done it? I would have liked to have had the first 40 or 50 people in the same office with me every single day. But COVID kind of broke that for everyone in our vintage.”

“I just had a really good network with people that owned businesses in Mexico. And then also just good old-fashioned hustle.”

— Deepak Chhugani

Watch Full Episode

Three Products, One Platform

Today, Nuvocargo is far from a Mexico-only play. Roughly 80-90% of what they do is within the US, with cross-border and Canadian freight rounding out the picture. They’ve expanded from full truckload to less-than-truckload and multiple equipment types. But the real strategic insight is in their product architecture — three offerings that mirror what the biggest winners in freight have built at scale, because “customers kind of demand of you to be a true partner for their supply chain.”

01 — Freight Brokerage

The most intuitive part of the business — and what many people assume is all Nuvocargo does. In reality, it’s about 40% of gross profit. They connect shippers with their network of 300,000+ carriers, using AI-powered procurement and negotiation to find competitive rates. The brokerage handles both domestic US freight and cross-border shipments.

For smaller customers spending under $5 million on freight annually, the relationship often starts here — “what’s your price?” — but quickly evolves. “If you can be consistently and structurally cheaper, at least for SMB customers, we have found that they just want to do everything with you,” Deepak notes. What begins as a transactional vendor relationship becomes an outsourced supply chain department.

02 — Customs Brokerage

A licensed, regulated service — think of it as being a specialized lawyer who prepares all the paperwork required to import or export goods through any US port. Nuvocargo is a US customs broker with CTPAT certification (a significant credential in the freight world that signals deep compliance capability). This product handles the regulatory complexity that makes cross-border logistics so daunting for most companies.

03 — Managed Transportation

This is the product Deepak is most excited about — and the one most aligned with where AI is taking the industry. The pitch to an enterprise customer spending $5-50 million annually on trucking: “Let me run all of that for you. I will obviate the need for you to hire 10 or 20 SaaS tools to coordinate everything. Your teams don’t have to learn how to use them. And we just charge you a small fee based on outcomes.”

The key distinction from brokerage: managed transportation isn’t about quoting a shipment from Dallas to New York. It’s about asking, “Do you even need to ship that? Could we combine it with 3 other shipments and reduce your spend overall?” Nuvocargo becomes the orchestration layer between the customer’s ERP and all their freight vendors — consuming purchase orders, optimizing routes, managing carriers. “That used to be with software and humans,” Deepak says. “Now it’s increasingly software, AI agents, and some humans overseeing what’s going on.”

The Hype Cycle Carousel

Deepak has a wry view of how the startup world has oscillated on business models over the past six years. “It’s been amusing to me to see how the startup world and the tech Twittersphere changes its mind,” he says, before cataloging the whiplash:

Through all of it, Nuvocargo kept building the same core thesis: charge customers to deliver outcomes. “We’ve done this like round trip back to the best model for customers and for value is what we’ve been doing all along,” Deepak says. “AI has just gotten so much better. We can actually offer so many agents, so many integrations right out of the box.”

“We lived 2019 to 2022, tech-enabled services, it’s the best thing in the world. Then 2022 to 2025, no, pure SaaS. And then it was like, I just want agents. I just want agents.”

— Deepak Chhugani

Why Most AI Agent Startups Aren’t Renewing

The AI agent renaissance in freight started with voice AI — automating carrier negotiations that industry veterans had sworn could never be automated. The use case was electrifying: AI systems that could call trucking companies, negotiate rates with drivers, and handle the messy, relationship-driven work of freight procurement. “This was one of those things that you would hear industry veterans saying, this cannot be automated with software. These Silicon Valley people are insane.”

And then it started working. Demos were phenomenal. Conference booths were packed. VCs got excited about freight again for the first time since the Convoy and Flexport era.

But Deepak has a different view from behind the curtain. “You really have to talk to people behind closed doors that are rolling this out. And they’ll tell you some successes they’re having, like, ‘Oh, now my inbound calls from carriers are run by these agents.’ But actually deploying this at scale where humans don’t need to monitor the agent anymore is incredibly difficult.”

The core problem is variation. Two freight brokers using the exact same TMS — the system of record in logistics — use it in fundamentally different ways. They have different business rules, different commission structures, different negotiation philosophies. Each customer requires intensive forward-deployed engineering and prompt engineering to get agents working. And even then, there’s an error rate that persists.

Worse, the failure mode is socially toxic. When agents break down, someone has to tell the customer that their team — not the AI — is the bottleneck. “It’s much easier to blame the vendor,” Deepak observes. “And that’s why I think a lot of them are not renewing. There’s a lot of controversy around whether this is going to work.”

The deeper strategic concern: if every broker adopts the same three or four agent vendors, where’s the differentiation? “It feels like it’s going to become another way that freight brokers commoditize themselves,” Deepak argues. “We all use the same 5 providers. It’s a little bit different, but not really.”

“If your AI strategy is just to work with a third-party vendor, all your competitors can hire the same thing. I don’t understand how that’s differentiation. It feels like it’s going to become another way that freight brokers commoditize themselves.”

— Deepak Chhugani

Watch Full Episode

Own the Outcome, Own the Agent

Nuvocargo’s approach inverts the typical agent model entirely. Instead of selling agents to freight brokers and hoping they figure out the integration, change management, and operational redesign — Nuvocargo sells the outcome. The pitch is disarmingly simple: “Just give me 10 shipments and I will show you the power of 12 AI agents and access to 300,000 trucking companies.”

Compare that to the typical agent startup pitch, which Deepak characterizes as: “Try my agent. Let me teach you my agent. Let me spend months on forward-deployed engineering.” Nuvocargo skips all of that because the agents run inside their own platform, NuvOS, where they control the entire workflow.

Each agent handles one narrow, deterministic task: a scheduling agent books pickup and delivery appointments. A document processing agent handles customs paperwork. A compliance screening agent vets carriers. Each step is binary — the agent either completed the task correctly or it didn’t. If it fails, a human steps in and unblocks it. If it succeeds, the next downstream step fires automatically.

“There’s not like one agent that runs all your freight,” Deepak explains. “There’s just too much potential for errors and variability. So we find that different customers might have different ways of doing different steps, and we’ll train our agent to figure that variability out. But they just do this one narrow thing.”

The customer never wrestles with prompt engineering, integration headaches, or agent management. They never have to retrain their teams to “manage agents instead of stale SaaS tools.” They just see results: lower spend, on-time delivery, fewer vendors, less work. And because the agents operate inside Nuvocargo’s platform across all customers, there’s a flywheel effect — each new customer’s carriers are often already onboarded from previous implementations, accelerating the next deployment.

“We Deprecated All External Views”

Between 2019 and 2023, during the tech-enabled services boom, Deepak admits he made a common mistake: spending too much time on the customer-facing dashboard — the one VCs would see in board meetings and customers would see in sales demos. The shiny external layer.

Then came a radical decision. Nuvocargo deprecated every external view of their software and went all-in on building NuvOS, their internal platform. “No one gives a shit about that data if it’s not reliable, correct, and timely,” Deepak says bluntly. “Especially in freight. No one cares.”

The insight is counterintuitive for any VC-backed startup: stop building the thing your investors want to see, and start building the thing that makes your operations bulletproof. When Deepak looks at shipment 4745, the data needs to be correct — tracking information present, documents right, appointments set. Getting that right internally makes exposing it externally trivial.

“A lot more of our roadmap has been internal,” he says. “And I think if you bonded with tech-enabled service founders or AI services founders — which I think is kind of the same thing at this point — they would tell you that the internal thing is much more important and getting it right at a smaller scale is much more important because iteration requires all this change management with the people in the operation.”

“No one gives a shit about that data if it’s not reliable, correct, and timely. Especially in freight. No one cares.”

— Deepak Chhugani

Internal vs. External AI Services

Nic Poulos offers a framework that resonates with Deepak: the distinction between internal and external AI services. An internal service is embedded within a customer’s operations — absorbing workflows, consuming first-party data, becoming part of how the business runs. An external service is just another vendor — maybe cheaper because AI automates costs on the back end, but not truly integrated.

Deepak’s response reveals how Nuvocargo thinks about this: “That’s one of the reasons we decided to move into managed transportation.” Freight brokerage alone can feel external — “what’s your price?” is the entire relationship. But managed transportation changes the dynamic. “When customers see what we’ve built, the first thing they ask is, ‘Can you integrate with my ERP?’”

He draws a parallel to Ramp: “When they started, I can’t tell you how many people were like, ‘It’s just a credit card. More cash back. It’s a commodity.’ Look at where they are today. It started that way, but then you build in more ways to add value to the point where they couldn’t imagine going off to the next guy just because the cashback is 10 bps higher.” The same dynamic is playing out in freight — but only for companies willing to go deep.

Pricing on Outcomes, Not Seats

Managed transportation pricing in freight isn’t something Nuvocargo invented — it’s a known model from multibillion-dollar incumbents. The standard: 2-4% of the customer’s total freight spend to manage their entire transportation operation. Then if the customer also wants access to Nuvocargo’s brokerage — the 300,000-carrier network — that’s a separate, complementary revenue stream.

Nuvocargo’s pitch isn’t “we invented a new pricing model.” It’s “you’re going to pay me the same as the managed transportation provider that started in 1980 or 2005, and you’re just going to get much better outcomes because of our modern technology stack.”

But Deepak’s most interesting insight on pricing is about restraint. He admits to learning the hard way during the ZIRP (zero interest rate policy) era: “No matter how amazing your product is, if you focus on squeezing every dollar of value upfront from the customer, first of all, you’re not going to win the business because they need to feel like it’s a great deal.”

The proof is in his numbers. One customer saw a 12% reduction in their multi-million dollar freight budget in year one. Deepak didn’t capture all that value — but the customer then quadrupled the amount of spend flowing through Nuvocargo and adopted additional products. “You have to believe you’re going to be able to add a lot of value, or your cost of giving them the service is going to keep going down — which we deeply believe and we see in our data.”

“I would not focus on top line. You’re just proving it’s a big market and a lot of people want to find the cheapest option. I don’t think that’s where value gets created.”

— Deepak Chhugani

Watch Full Episode

From “AI Is the Next Crypto” to “You Guys Need to Chill”

The internal transformation at Nuvocargo mirrors what many companies are going through — but Deepak is unusually candid about how it unfolded. Early on, respected people in his R&D org and product operations talked about AI “like the next crypto or VR, like a bubble, like something that doesn’t add value.”

Fast forward to the company’s most recent executive offsite: “I swear I had to tell them, ‘You guys got to chill out. This is too much AI. All we’re talking about is AI.’” For Deepak, that whiplash illustrates how radical the transformation has been.

His approach to driving that change was deliberately grassroots. Step one: be the annoying CEO. “I was that annoying guy spamming people with, ‘Check out this AI startup, check out this VC talking about how AI is going to change everything.’” He knows it was annoying — he draws a parallel to VCs who forward their own content to founders. But he didn’t see another way.

Step two was more tactical. Leveraging his YC network, Deepak opened the doors to any AI startup that wanted to pitch Nuvocargo. “Anyone who wanted to pitch us, you can pitch me. I don’t care if it’s 5 hours of my day. You can pitch me because you’re spending all your time with the new tools and I want to see.” He’d loop in team members from operations and engineering. “60% of the time they would say it was garbage. But the 30 or 40% started being like, ‘Holy shit, I can’t believe that this is possible.’”

The journey wasn’t smooth. They churned through multiple AI vendors — some overpricing, some underdelivering. They tried voice AI providers, email AI systems, carrier negotiation tools. But gradually, the cultural momentum shifted. “It felt like a year of being behind, and then now fast forward, I have people dropping in our inbox telling me, ‘I want to join your company because my company feels 4 years behind what you guys are even publishing.’”

Selling the Story of the Real Economy

How do you recruit top technical talent to a freight logistics company when they could go work at OpenAI? Deepak credits Ryan Peterson (Flexport) and Dan Lewis (Convoy) for showing the industry how to tell the story differently.

“You could tell a story of, ‘Hey guys, come join me, I know I’m not as sexy as OpenAI.’ That’s a defeatist story. We never had that story.” Instead, Nuvocargo’s pitch leans into the tangibility of freight: “Look at all these AI wrappers, 57 competitors building the same thing. You’re doing something in the real economy. I have customers moving stuff into data centers — plastics, toilets, the real world. Every little thing in the room you’re sitting in was moved in trucks.”

It’s working. Deepak says people have left Nuvocargo for “sexy companies” and come back: “They’re like, ‘I don’t even know what I’m doing. I’m working on marketing software — for what?’” Trade as a narrative — the thing that connected civilizations, that moves the real economy — turns out to be more compelling than another SaaS dashboard, at least for the right kind of builder.

“You take something that sounds unsexy and you have a story that’s completely factually true and it gets people much more excited about the significance of what you’re doing.”

— Deepak Chhugani

The Three Metrics That Actually Matter

When asked what metrics AI service founders should obsess over, Deepak offers a clear framework — one born from making the opposite mistakes during the ZIRP era.

First: Can you structurally reduce your cost to serve? Not just be cheaper today, but fundamentally transform the cost structure of delivering your service through technology and AI. “If you can make it from a service business to almost like an API or a consumption-based business, it becomes inherently more valuable. Businesses like that trade at much higher multiples than human-heavy service businesses.”

Second: Does cheaper cost to serve produce better outcomes? This is where AI becomes transformative rather than just cost-cutting. Nuvocargo can interact with 300,000 carriers, negotiate, and screen for compliance in minutes. “Even the best carrier sales rep in our industry cannot do that.” The promise isn’t just “same thing, cheaper” — it’s “things that were literally impossible before.”

Third: Can you price based on outcomes? Not seats, not usage, not tools — outcomes. Did the shipment cost less? Did it arrive on time? Is the customer’s life simpler? “If you measure those two things and then find a way to price based on outcomes — those are the three things I would be thinking about if I was starting from scratch. And they’re all hard. But it’s meant to be hard, so you have real value at the end. Not just raise your Series A and then kind of be fucked.”

Key Takeaways

From the Episode:

“Just give me some shipments and I will show you the power of 12 AI agents.”

Listen to the full conversation between Deepak Chhugani and Nic Poulos on the Verticals Podcast, Episode 21. If you’re moving freight — or know someone who is — visit nuvocargo.com and try it with 10 shipments or less.

This post was written based on Episode 21 of the Verticals podcast, a show about vertical AI and industry-specific technology companies.

Watch Full Episode

Have a product or service that would be great for our audience of vertical SaaS founders/operators/investors? Reply to this email or shoot us a note at ls@lukesophinos.com

Today’s newsletter is sponsored by Xplor Pay. In their latest case study, they show how a POS provider embedded payments using PayFac as a Service without taking on the burden of compliance, risk, or merchant onboarding.

Learn how white-labeled payments, streamlined setup, and flat-rate pricing helped the platform improve the experience while unlocking new recurring revenue. Read the full case study.

Alright, let’s get to it…

The Only Non-BS Way to Decide Which AI Agents to Add to Your vSaaS

Everyone wants “AI agents” in their product right now.

Most teams ship a chat box, call it an agent, and then… nothing changes. Usage is cute. Retention doesn’t move. Nobody’s day actually gets easier.

Build your agent roadmap by ranking “automation level” per role (not by vibes)

The first step is actually building a standard org chart for the customers you serve.

Map out EVERY single job/role at your ICP. This should include the Title, the full job description, the compensation, and the specific milestones/quantitative goals they are responsible for achieving.

1) Build your AI Agent roadmap by ranking the level of automation you can achieve per role

The question is not “which role can we automate end-to-end?”

The question is: for every role that touches your product, what level of automation can you credibly deliver—and how fast can you move that role up the automation ladder? Then you force-rank the roadmap based on those deltas.

This works especially well in vertical SaaS because vertical agents can actually bake in domain rules, compliance constraints, and industry jargon that generic models miss. That’s the point of “vertical AI agents.”

Step 1: Map the customer org chart (the workflow lives in the handoffs)

Start by mapping the org chart for your ICP account type, not just “the user.”

You’re looking for:

• who generates inputs

• who transforms inputs into records

• who approves/rejects

• who gets blamed when it’s wrong

If you sell into construction, that might be Supers, PMs, VDC, PE’s, owners. If you sell into logistics, it’s dispatch, ops, billing, customer service. If you sell into healthcare ops, it’s front desk, MA, coding, prior auth, billing.

Why this matters: agents don’t create value in isolation. They create value when they reduce coordination cost across roles (handoffs, “what’s true?”, “who owns this?”, “where’s the proof?”).

Step 2: Write a job description for every role (this is where agents hide)

Yes. Write the job descriptions.

Not HR-perfect. Operator-real.

For each role, capture:

• what they’re accountable for (KPIs)

• what they do daily/weekly

• what decisions they own

• what artifacts they produce (reports, logs, packets, approvals)

• what failure modes create rework, disputes, delays, or compliance risk

This turns “we should add agents” into a finite set of workflows with definitions of done.

Step 3: Put each role on an “Automation Ladder” (LoA) and score it

Here’s the ladder (use this exact thing in a spreadsheet):

LoA 0 — Manual
The role is meetings, tribal knowledge, and ad-hoc work.

LoA 1 — Retrieve & explain (read-only)
Agent can answer questions, find context, summarize history.

LoA 2 — Draft artifacts
Agent drafts real deliverables: emails, daily logs, status reports, inspection packets—with citations back to source records.

LoA 3 — Recommend decisions
Agent flags exceptions, recommends approve/reject/prioritize, explains why.

LoA 4 — Execute with guardrails
Agent updates records, routes work, requests missing info, triggers workflows (with policies/approvals).

LoA 5 — Closed-loop outcomes
Agent completes the workflow end-to-end and handles most exceptions.

Now score each role on two dimensions:

1. Ceiling: what LoA is realistically achievable in 12 months given your data + permissions + integrations?

2. Speed: how fast can you move the role up 1–2 levels?

If you do this honestly, you’ll find something surprising: a lot of roles can jump from LoA 0 → LoA 2 quickly (drafting artifacts is an immediate win), while only a few can safely get to LoA 4 early (execution needs guardrails + trust).

Step 4: Force-rank your agent roadmap using “LoA Delta × Frequency × Blast Radius”

This is the “stop arguing” formula.

For each role:

• Current LoA

• Target LoA in 90 days

• Target LoA in 12 months

• LoA Delta

• Frequency (daily, weekly, monthly)

• Blast Radius (how many downstream roles depend on the output)

Then force-rank your roadmap by:

Roadmap Priority Score = (LoA Delta) × (Frequency) × (Blast Radius)

That’s how you avoid building “cool agents” and instead build the ones that become operationally unavoidable.

Step 5: Turn your top 3 roles into “Agent Job Stories” (not feature lists)

For your top 3 roles, write job stories like:

“When it’s Thursday and I need to approve invoices, I want an evidence-backed packet that shows what’s complete, what’s missing, and what’s risky—so I can approve or reject quickly.”

“When an inbound request hits the queue, I want it classified, deduped, enriched, routed, and tracked—so nothing slips.”

Notice what’s consistent: a trigger, an outcome, and a decision.

Step 6: Define the “Agent Contract” (this is what makes it sellable)

Every agent you ship should have an explicit contract:

Inputs
Where it pulls data from (your SoR objects + integrations).

Outputs
What it produces (and what “done” means).

Citations
Links back to source records (this is trust).

Guardrails
What it will never do automatically.

Write-backs
Which objects it creates/updates in your system of record.

This is the difference between “AI demo” and “workflow product.”

Step 7: Ship the trust ladder (Draft → Recommend → Execute)

Here’s the rollout pattern that works in real vertical ops:

1. Draft (LoA 2): generate artifacts with citations

2. Recommend (LoA 3): propose decisions + highlight exceptions

3. Execute (LoA 4): take actions with policies + approvals

4. Close-loop (LoA 5): only when your exception handling is real

If you try to skip to execution, you’ll get blocked by risk, permissions, and “what if it’s wrong?” politics.

Step 8: Examples of actual jobs (and the agents that map cleanly)

To make this real, here are actual job archetypes that show up in most operational verticals, and why they score high on automation potential:

1) Intake & Triage Coordinator
Their day is routing, deduping, filling missing fields, chasing attachments, and assigning ownership.
Agents here jump fast because “done” is crisp: a clean record exists, assigned, with next steps.

2) Status Translator (the “nobody trusts the truth” role)
They pull from 3–7 systems and produce the weekly narrative for leadership/customers.
Agents win here at LoA 2–3 because drafting + exception flagging saves real time and reduces surprise.

3) Evidence Packet Builder
They compile proof for approvals, compliance, billing, audits, disputes, inspections.
Agents here deliver immediate ROI at LoA 2: a source-linked packet + what’s missing + what’s risky.

4) Scheduler / Dispatcher
They coordinate people/assets against constraints (availability, location, priority, SLAs).
This often starts at LoA 3 (recommendations) and grows into LoA 4 execution with guardrails.

5) Compliance / Audit Coordinator
They track obligations, deadlines, control evidence, and audit trails.
Agents do well here because the work is artifact-heavy and rule-heavy (vertical knowledge matters).

Step 9: Your “copy/paste” worksheet

Create a table with columns:

• Role

• KPI / accountability

• Recurring artifacts produced

• Current LoA

• Target LoA (90 days)

• Target LoA (12 months)

• LoA Delta

• Frequency

• Blast Radius

• Top 3 integrations needed

• Guardrails needed

• Metric that proves ROI

Fill this out for 10–20 roles and your roadmap will basically write itself.

Step 10: Success metrics (what to measure so you don’t fool yourself)

If you want to know whether the agent is real, track:

• Cycle time reduction (time-to-close, time-to-approve, time-to-route)

• Artifact completion rate (reports/logs/packets created on time)

• Exception resolution rate (how many issues caught early vs late)

• Adoption in the role (not overall MAU—role-specific penetration)

• Downstream impact (fewer disputes, fewer escalations, fewer missed steps)

If you can’t tie your agent to one of these, it’s a demo.

Let me know how this goes. I’ve been doing this with a few companies and it has been so eye opening for us…

The Story of Drone Deploy— from “drone mapping app” to “agentic site intelligence platform”

DroneDeploy is one of those companies that’s easy to misunderstand if you freeze them in time.

If you met them in 2016–2018, you’d call them “drone mapping software.”

If you look at them now, the better description is: a reality capture system-of-record with an AI agent layer that turns messy site conditions into decision-grade outputs (progress, safety, inspection). They literally describe it as “bring the site to you” using drones, robots, 360 cameras, and AI agents—and say they’re trusted by 5,000+ enterprises.

Here’s how they got there, and what you should steal.

The origin story: three founders, one obvious missing layer (software)

DroneDeploy was founded in 2013 at AngelPad by Mike Winn (CEO), Jono Millin (CPO), and Nicholas Pilkington (CTO).

What they saw early was simple: drones were getting adopted, but the “value” wasn’t the drone. It was what the drone produced: a map, a model, a repeatable way to see what changed.

So they built the software layer to plan flights, capture data, and turn it into usable outputs for industries like agriculture, construction, and insurance.

At the time, the company leaned into a SaaS model with a free tier and paid plans (CNBC noted premium plans starting at $99/month back then).

Early funding: Scale leads Series B as commercial drones open up

In 2016, DroneDeploy raised a $20M Series B led by Scale Venture Partners (with High Alpha participating), bringing total funding at the time to $31M.

What’s notable is why this round happened when it did: commercial drone regulations were tightening into something usable, and the “software control plane” became the obvious pick-and-shovel.

TechCrunch framed DroneDeploy as a platform used across a ton of industries to plan flights and produce detailed maps/3D models.

The strategy shift: “beyond drones” and a multi-modal product suite

This is where the story gets very vSaaS.

DroneDeploy didn’t stay trapped as “a drone app.” They expanded the capture modalities and built toward a unified record of site reality.

You can see the product suite turning into a platform:

1) Construction system-of-record motion
On the construction page, they position the platform as “one platform for all your site and photo documentation.” They explicitly talk about automatically organizing photos across the lifecycle, sharing annotated reports and 360 walkthroughs, and improving safety by inspecting hard-to-reach areas.

They also highlight integrations with Procore and Autodesk Forge (drawing sync, RFIs/observations, embedded app experience, photo archive), plus exports to tools like Navisworks and Revizto.

That’s important: they’re not trying to rip out the stack. They’re trying to become the “visual truth layer” the stack can reference.

A quote on the construction page captures the SoR ambition perfectly: “We finally have a single project management tool for all site documentation in one place.” (Austin Lay, STO Group).

2) Robotics and automation as the capture engine
Their Robotics product messaging is basically: schedule missions (docked drones + ground robots), automate collection + upload, then let AI add context. They frame it as “three solutions, one platform,” including Dock Automation and robotic walkthrough/inspection workflows.

3) Acquisition-led consolidation
In 2022, DroneDeploy announced it would acquire StructionSite to unify aerial + ground capture into one platform (interior + exterior).

They backed that strategy with real scale metrics: they said they’d mapped 300+ million acres and 1.7 million sites from 7.9 million flights and served customers across 190 countries.

And StructionSite brought its own gravitational pull: DroneDeploy said StructionSite customers captured $190B of construction volume, with 3x annual growth in active projects, and “over 500 builders” using the product.

Say whaaaaaaaat. That’s not “nice-to-have documentation.” That’s the beginning of a category: record the site, then run the business off the record.

Funding milestones: from “category leader” to “break-even + strategic AI/robotics funding”

DroneDeploy also has a long arc of funding that maps to their strategy evolution:

• Series D (2019): They announced a $35M Series D led by Bessemer, bringing total funding to $90M.

• Series E (2021): Energize noted it co-led DroneDeploy’s $50M Series E with AirTree, and said total capital raised hit $142M. Energize explicitly called out that the capital would help expand beyond aerial capture and pursue acquisitions.

• 2025: DroneDeploy said it reached break-even and raised a $15M strategic investment to accelerate AI and robotics (Progress AI, Safety AI, autonomous ground robots), with investors like Emergence, Scale, Airtree, Bessemer, and Uncork “doubling down.”

The break-even line matters. It’s a signal they can fund a long roadmap (robots + AI) without being forced into short-term, growth-at-all-costs decisions.

Customers (real examples) and why they buy

DroneDeploy sells into industries that have the same core pain: field reality is expensive to capture, inconsistent, and hard to share—so decisions get made on partial truth.

A few customer examples they themselves highlight:

Turner Construction (robotic walkthroughs):
Turner uses DroneDeploy’s robotics solution + Boston Dynamics Spot to run overnight site capture. Turner’s Reality Capture Manager quoted saving 4–5 man hours per 15-minute automated mission.

Layton Construction (portfolio visibility + Progress AI direction):
In their break-even + $15M announcement, DroneDeploy includes a quote from Layton’s VP of VDC about getting accurate visibility in minutes and believing in the automation roadmap across drones, robotics, and AI—especially for complex projects like data centers.

Wharton-Smith (Progress AI job-to-be-done):
In the Progress AI launch, they quote multiple Wharton-Smith users describing Progress AI as “like having an extra superintendent,” and one person supporting “over 90 projects” who can now check progress without stepping on-site.

The current product suite: unified capture + unified record + AI agents

Here’s the product suite overview as of their 2025–2026 narrative (in plain operator terms):

Reality capture platform (core): unify drones, 360 cameras, and robots into one searchable record “by date and location,” so teams can reference “what’s actually happening” without being on-site.

Progress AI: generates structured progress outputs quickly—within minutes after capture—with items like percent complete by location, trade status, and installed work across 80+ trade types.

Safety AI + Inspection AI: At Horizons 2025, DroneDeploy described three operational agents—Progress, Safety, and Inspection AI—positioning them as automating visual intelligence across construction and industrial sites.

Robotics platform: schedule autonomous docked drone flights and robotic walkthroughs/inspections that automatically upload data into the platform.

That’s the full loop: capture → store → analyze → act.

The wedge (what to steal): “make reality queryable, then make decisions automatic.”

The DroneDeploy wedge is not “drones.” It’s not even “3D models.”

It’s this sequence:

1. Make site reality easy to capture (more modalities over time: aerial + ground + robots)

2. Make it trustworthy and searchable (date/location organization, unified platform)

3. Turn that record into structured outputs (Progress AI outputs, safety/inspection agents)

4. Expand into more jobs and more budgets (construction + energy + infrastructure + agriculture; and “trusted by 5,000+ enterprises”)

That’s vertical SaaS compounding. The record gets better, the models get better, the outputs get more trusted, and the platform becomes harder to replace.

Chris Hladczuk sat down with Nic and Luke Sophinos to unpack how Hanover Park went from zero to fifteen billion in assets under administration—and why AI-native services are eating B2B SaaS alive.

Watch & Subscribe Here

Vertical Titan: A Maniac On A Mission

Chris Hladczuk is the kind of founder who makes you reconsider what ambition looks like in practice. Before Hanover Park, he was deep in the trenches of financial services—not as a technologist looking in from the outside, but as someone who understood the industry’s rhythms, its pain points, and its deeply entrenched trust dynamics. He saw what most people saw: an industry dominated by legacy players, allergic to change, and held together by relationships built over decades. But where others saw an impenetrable fortress, Chris saw a category ripe for reinvention.

The insight wasn’t that fund administration needed better software. Plenty of people had tried that. The insight was that fund administration needed a fundamentally different kind of company—one that combined the service intensity of a white-shoe accounting firm with the technological ambition of an AI-native startup. Hanover Park was built on the conviction that if you could automate the invisible, grueling, error-prone back-office work with AI agents, you could redeploy every human toward high-caliber advisory work. The result: a team Chris calls “the Navy SEALs of fund accounting.”

In twelve months, Hanover Park went from zero to fifteen billion dollars of assets under administration. They raised a $27 million Series A from Emergence Capital. They have over 10,000 LPs on the platform. And Chris believes they’re just getting started—because the real unlock isn’t the product. It’s the model: sell outcomes, not software. Charge basis points, not subscriptions. And never, ever say the word “AI” to a CFO.

When Nic opened the conversation by asking about the trust problem—the deep-seated human need to have a person you can call when you’re frazzled, when the numbers don’t add up, when you need someone who has your back—Chris didn’t dodge it. He leaned into it. Because the trust problem is exactly the problem Hanover Park was designed to solve, not by eliminating trust, but by earning more of it.

The Trust Problem. Why you pay your accountant isn’t about the math.

Nic framed it perfectly: “Why you pay your accountant, your private wealth advisor, your tax guy—yes, of course you want it to be accurate. But you also want somebody you trust. You can call up Joe or Kim and they’ve got your back.” In fund administration, this dynamic is amplified by the sheer complexity of the work and the magnitude of the stakes. A fund admin’s mistake doesn’t just mean a wrong number—it can mean failed LP reporting, botched audits, or a CFO losing their job.

Chris’s response to this was surprisingly old-school for someone building an AI company: he didn’t talk about replacing the relationship. He talked about making it better. “Automate the work you don’t see,” he said, “so we can provide high-level advisory work and white-glove service for the things you do see.” The idea is deceptively simple: by deploying AI agents to handle the core transactional layer—cash reconciliation, document parsing, KPI extraction—you free up every human on the account to be a high-caliber CPA focused on complex, edge-case advisory work.

The structural advantage this creates is profound. Because Hanover Park doesn’t need an army of humans clicking buttons to reconcile cash, they can hire fewer, better people—people they can pay more, people they can give equity to. The result is a team of what Chris calls “the Navy SEALs of fund accounting.” It’s a phrase he’s quite proud of. “I don’t think those four words have ever been said together in the same sentence,” he laughed.

“It feels like I have 25 people on my account when they have two.”

— Hanover Park customer

That customer quote isn’t marketing copy. It’s the lived experience of what happens when you combine elite human talent with AI agents that handle the mundane. And it gets at a subtlety that most AI companies miss: the goal isn’t to replace the human. It’s to make the human so much more effective that the customer feels like they have an entire team dedicated to them—because, in a way, they do.

Nic pushed back on this, drawing from his own experience as a fund manager. “When these models try to scale and they get rid of the Joe personally, and you switch it out with Euclid at whatever, it becomes impersonal and context is lost.” Chris agreed—which is why Hanover Park assigns a dedicated team to every account. No rotating cast. No faceless support tickets. Your team is your team.

Watch & Subscribe Here

Today’s episode is brought to you by Parafin - the leading embedded capital partner.

• 4.8 star rating on Trustpilot, with over 500 reviews [source]

• 71% CSAT rating across all partners

• 84 NPS

• Serving top Fortune 500 companies, including Amazon, Walmart, and DoorDash.

AI prepare. Human review.

When Nic asked about the tech stack—specifically, how Hanover Park handles hallucination risk in an industry with zero fault tolerance—Chris laid out his operating paradigm in four words: “AI prepare, human review.”

Here’s how it works in practice. A fund manager—let’s call him Nic, because that’s literally who Chris used as an example—forwards a legal invoice to Hanover Park. At a traditional fund admin, someone would force Nic to log into a portal, fill out forms, and wait. At Hanover Park, the email itself is the interface. An AI agent reads the email, creates tasks, proposes journal entries, and executes against the ERP. A human reviews and approves—the whole thing takes minutes.

“Literally someone called me who switched over from a previous fund admin,” Chris recalled. “They said, ‘How did that happen in six minutes? That was crazy.’ But a human literally had everything queued and all the human did was approve and reply back automatically.” The customer gets the speed of AI. Hanover Park gets the accuracy of human review. Everyone wins.

This matters because in financial services, accuracy always trumps speed. Chris was explicit about this: “The worst pitch ever to a CFO is ‘trust me, it’s AI.’ That’s not a good pitch. I don’t actually say the word AI very much.” Instead, he lets outcomes speak for themselves. When a CFO sees their data updated in real-time instead of waiting weeks for a quarterly reconciliation, they don’t need to know there’s an AI agent behind it. They just know the service is exceptional.

Be where your customers are.

A recurring theme throughout the conversation was Chris’s disdain for the “2018 SaaS paradigm”—forcing customers into a portal to click seven buttons. Instead, Hanover Park builds its products around how CFOs and fund managers actually work: email, board decks, and conversation.

✉️ Email Agents

Forward a bill, delegate like you’re delegating to a human. The AI agent reads the email, creates tasks, proposes journal entries, and executes on top of Hanover Park’s ERP. A human reviews and approves in minutes. One customer asked, ‘How did that happen in six minutes?’

📊 KPI Reconciliation

Send a board deck or founder update—any format. Hanover Park’s KPI reconciliation agent pulls out context, extracts metrics, and cross-references against the ERP and the old-school SaaS forms. Most customers don’t even fill out KPI forms anymore. They just send an email.

🔍 Portfolio Intelligence

Every deal generates accounting data. Hanover Park weaponizes it—AI-parsed legal docs, cap tables, board decks, and KPIs—bundled free alongside core fund admin. It’s three tools competitors sell separately, delivered better because they share the same source of truth.

🔗 MCP Server

Every piece of data in your firm’s history, accessible through a model context protocol. Generate any customized PDF report on demand. ‘No other fund admin has even considered this,’ Chris said, ‘because the industry is so old school they don’t call it a product—they call it a system.’

Watch & Subscribe Here

B2B SaaS is dead.

In September 2024, Chris went on Turner Novak’s podcast, The PLN, and declared B2B SaaS dead. “Everyone told me I was an idiot,” he recalled. “Now, with Claude and Anthropic, it seems like I’m not an idiot. This actually makes a lot of sense.”

The argument is structural, and it’s worth unpacking. If building software becomes trivially easy—and it is becoming trivially easy—then the application layer commoditizes. What Chris calls “Opus 4.6 to Opus 5.6” will eventually be able to handle the execution layer that many vertical SaaS companies rely on today. The question he asks is: what are the model providers not going to do?

His answer: OpenAI and Anthropic are not going to build a fund administration services company that does accounting for Nic. They’re not going to hire CPAs, stand up an ERP, handle LP reporting, and fly to a CFO’s office at midnight when something goes wrong. Even if the AI can do the core work, no one at those companies wants to be a services business that handles fund accounting. That’s the moat.

“Sell outcome, meaning do accounting—not sell software to accounting firms,” Chris summarized. “I think that’s more interesting and has less potential disruption risk.” The business model reflects this conviction. Hanover Park doesn’t charge a SaaS subscription. They charge basis points on AUM. As Nic pointed out, this gives them revenue quality that mirrors traditional subscription businesses—with the added benefit that your worst outcome as a closed-end fund is a ten-year relationship.

“If Brett Taylor is thinking that the system of record is less important—it’s what the agent context has—that pushes my thinking on how we need to be intellectually honest about this.”

— Chris Hladczuk

When Luke asked whether software companies might eventually become AI-native services businesses themselves, Chris was skeptical. “The entire way in which you approach the market is different,” he said. “I have fund accountants sitting ten yards from me in New York City. Fund accountants have never joined a startup before. Period. That is a very different recruiting function, operations function. It’s so different from having a bunch of engineers, salespeople, and designers in a room.” The cultural barrier, he argued, is underrated. Software companies becoming services companies isn’t just a business model shift—it’s a DNA transplant.

Mr. Bundle.

Most investment firms buy five to seven standalone tools: portfolio management, portfolio monitoring, valuations, tax, LP portal, fund modeling. None of them are great. None of them talk to each other. And none of them share the core accounting data that makes them useful.

Hanover Park bundles all of them into core fund administration, powered by the same source-of-truth ERP. It’s not a freemium play. Chris will never sell standalone portfolio tools—a stance that Luke actively pushed back on during the episode. “Couldn’t it be valuable to have every single fund on Hanover Park and then cross-sell them over time?”

Chris’s response was categorical: “I am Mr. Bundling. I literally hate the word unbundle. I will literally never sell the standalone product of Portfolio Tools.” His reasoning is specific to the industry: the reason other standalone portfolio tools aren’t great is precisely because they lack the core source-of-truth data that lives in the accounting system. “When you unbundle, the products are actually worse. So now I’m going to deliver you a bad portfolio experience? That’s not going to be great.”

The wedge product—the thing that gets funds in the door—is the portfolio intelligence tooling that AI-parses legal docs, cap tables, and board decks. But the wedge only works because it’s bundled with the core fund admin. Separate it, and you lose the data advantage that makes it compelling.

“I literally hate the word unbundle. When you unbundle, the products are actually worse.”

— Chris Hladczuk

Always get on the plane.

Switching fund admins is, as Chris put it, “asking someone to get married on the first date.” When it’s one or two funds, it’s not that big of a deal. But when you’re ripping out seven funds with 22 entities, 37 sub-structures, and 10 SPVs on top, the CFO has to look you in the eyes and decide: do I trust this person with my job?

Chris’s approach to this is radically personal. “I literally, at the drop of a hat, will get a text message from a CFO being like, ‘Hey, I don’t know about this.’ I’ll call them immediately, book a flight, and be there tomorrow morning. Then we’ll lock ourselves in their office for four hours and get to work.” He coins it as “always get on the plane”—a phrase he claims to have invented on Twitter, which has since been widely adopted.

Luke pushed him on sustainability: “You’re the maniac on the mission. You tweet about sleeping at their offices. Tell the founders listening what this actually takes.” Chris was candid: “We need to get out of founder-led sales. I’m on seven jobs right now. We’re a little too big for this, honestly. But I like staying close to it.” The intimacy isn’t just a phase—it’s a signal. If the founder isn’t willing to get on a flight for the most important vendor decision a CFO will make, why would the CFO trust the company for the next ten years?

The zero-to-one story was predictably unglamorous. The seed investor became a customer. A handful of early believers took a chance. “Until probably five billion, it was still very uphill,” Chris admitted. “Very challenging. So much embedded trust, LP institutional complexity.” The breakthrough wasn’t a single moment—it was a compound of relentless customer obsession. The three elements of a raving fan, as Chris defines them: incredible onboarding, 99.9% accurate and fast delivery, and jaw-dropping AI-native product experiences. “I don’t think you can survive on two without three.”

“I’d rather you not become a customer than become an unhappy customer in four months.”

— Chris Hladczuk

The hardest part is saying no.

Hanover Park doesn’t take funds under $100 million. Chris is turning away customers “left and right” who aren’t the right fit. This is counterintuitive for a company in growth mode—most startups would take every dollar they can get—but Chris sees customer selection as an existential discipline.

Nic compared it to the CISO problem: “When you’re doing your job really well, I probably don’t think about you that much. When you’re doing your job poorly, I hate you passionately.” Chris agreed, extending the metaphor: “It’s like offensive line in football. No one cares about you until someone gets sacked.”

His North Star isn’t NPS. It’s whether every single customer would enthusiastically refer Hanover Park—what he calls a “raving fan.” “That’s insanely hard in our space, by the way. Making sure someone’s a raving fan is very challenging. That’s why I’ve been super disciplined.” Where a typical vertical software company might optimize for volume—sell it a million times, have a million customers—Chris is optimizing for love. It’s a different game entirely.

Live in healthy fear.

When Nic asked about switching costs—a topic they’d explored in previous episodes—Chris reached for Jeff Bezos: “I live in constant fear of my customers. I literally told our team, I don’t even care if we never get another sale. I just want every customer to be a raving fan.”

It’s not entirely exaggeration. Hanover Park has 10,000 LPs on its platform today, including major institutions. The LP portal creates lock-in—are you going to tell Harvard’s endowment they need to create a new login next month? But Chris doesn’t lean on that as defensibility. He assumes his customers can leave instantly and builds as if they might.

Luke asked if he’d felt any churn impact from Claude Code and similar tools. “No,” Chris said, but not because he’s complacent. “It’s different. It’s software plus services plus output plus accounting delivery. There’s a lot there.” What excites him more is the offensive potential: if AI can supercharge migration—taking twenty years of terribly-kept complex financial data and moving it cleanly to a new system—then the constraint on growth isn’t sales anymore. It’s physics.

“Assume my customer can leave me instantly. Then build something so good they never want to.”

— Chris Hladczuk

Own the ledger.

Every smart investor told Chris not to build a general ledger from scratch for the most complex fund accounting in the world. “You’re an idiot. Don’t do this. You’re dumb.” He did it anyway.

His reasoning comes down to a single insight: AI agents are just clicking buttons on top of the ERP. If the ERP can’t handle the core action, it doesn’t matter how smart the agent is. Without a foundational general ledger built from scratch, “you’re just a UI on top of somebody else’s system.” He’s dismissive of competitors who white-label another ERP and layer AI on top. “I don’t think that works.”

Chris acknowledged the tension with Brett Taylor’s thinking at Sierra—the idea that agents can talk to any system of record, making the system itself less important. “He’s a thousand times better of an operator than I am. So what do I know?” But he’s not ready to concede the point. In his industry, the ERP isn’t just a database—it’s the source of truth that makes every other product in the bundle better. Lose that, and you lose the data advantage that separates Hanover Park from every other standalone tool.

Key milestones.

Sep 2024

“B2B SaaS is dead” thesis

Chris goes on Turner Novak’s podcast. Everyone says he’s wrong.

Q4 2024

Rapid onboarding begins

First wave of enterprise fund admin clients migrated.

Early 2025

$27M Series A

Led by Emergence Capital. Team scales to elite CPAs + AI engineers.

Mid 2025

$15B AUM milestone

10,000+ LPs on platform. From one to fifteen billion in 12 months.

Next

One-click migration

Long-running AI agents to process decades of complex financial data.

Looking Ahead. Achieving “One-click migration.”

When Nic asked his final rapid-fire question—what’s the biggest limitation of AI for your business, and what would you change overnight?—Chris didn’t hesitate: long-running, long-horizon agents.

The context is specific. Migration is the single biggest constraint on Hanover Park’s growth. Chris frames it the way Elon Musk frames manufacturing bottlenecks: what is the limiting factor in the business? “If I could click one button and migrate, we would not have 15 billion. We’d have a hundred billion.” The challenge is that you’re asking an AI to take twenty years of complex, poorly-maintained financial data from legacy systems and produce a clean, accurate migration. No one has cracked this yet.

But Chris is betting on a three-to-six month horizon. The frontier of long-running agents—systems that can process complex data over hours or days, maintaining context and accuracy—could fundamentally change the unit economics of customer acquisition. “I’ve been saying this for literally a year. My team laughs at me. But I’m building a one-click migration future.”

If he’s right, the fund admin industry won’t just be disrupted. It’ll be compressed. Decades of accumulated switching costs evaporated by an agent that can run for a week and come back with a clean general ledger. It’s not science fiction. It’s a roadmap.

“Take all the most complex financial data in the world, come back to me a week later. That as an idea is the frontier.”

— Chris Hladczuk

For Founders: The canvassing playbook.

Nic asked Chris what advice he’d give to a new founder thinking about entering a category dominated by a legacy services provider. Chris’s answer was refreshingly unscalable: “The only antidote is the first hundred days of the company. Obsessively talking to unlimited people in the ecosystem. There is no antidote to talking to a thousand CFOs.”

He broke his process into three layers of canvassing. First, customer canvassing: stress-testing every assumption about the market by talking to hundreds of buyers. Second, failure canvassing: talking to people who tried something similar and understanding why it didn’t work. Third, competitive canvassing: mapping every player, their strengths, their gaps, and where you can build something more compelling.

“I had 65 pages on every single competitor, every single thing they had, every single thing I thought I could do differently,” Chris said. “That’s what got me to conviction. Just like an investor. But I only have one thing to work on for twenty years.”

What Chris believes.

Don’t say AI.

The worst pitch to a CFO is ‘trust me, it’s AI.’ Chris rarely uses the word with customers. The product speaks through outcomes, not buzzwords.

Own the ERP.

Without a foundational general ledger built from scratch, you’re just a UI on top of someone else’s system. Every smart investor said don’t. Chris did it anyway.

Customer selection is survival.

Mid-market and enterprise only. No sub-$100M funds. Turning away wrong-fit customers is a feature, not a weakness.

Raving fans over growth.

North star isn’t NPS. It’s whether every single customer would enthusiastically refer you. ‘That’s insanely hard in our space, by the way.’

Services aren’t a dirty word.

AI-native services is the disruption. Software companies won’t easily become services companies—the cultural DNA is too different.

Live in healthy fear.

Assume your customer can leave instantly. Build something so good they never want to. ‘I don’t even care if we never get another sale.’

Always get on the plane.

If a CFO texts with a concern, call immediately. Book a flight. Be in their office tomorrow morning. Four hours locked in a room until it’s sorted.

High ego for vision, low ego for execution.

‘You have to have some sort of ego to think we can build something that big, but low ego in terms of day-to-day execution.’

The Takeaway

Nic closed the episode by telling Chris he was sold. Luke asked if he’d gotten the deal. Chris laughed: “I show you the product and it’s over. I don’t know why we haven’t even gotten there yet.”

That line captures something essential about Hanover Park and about Chris as a founder. He spent an hour talking about trust, culture, obsessive customer service, sleeping in offices, and flying across the country at midnight—and never once showed the product. Because the product isn’t the pitch. The pitch is: we will care about your fund more than anyone else in the world, and we’ll have the AI infrastructure to prove it.

In an industry where the old guard calls engineering “IT” and calls their platform “a system,” Hanover Park is building something that feels inevitable in retrospect but impossible in the moment. Fifteen billion in twelve months suggests it’s working. The hundred-billion target suggests it’s just getting started.

Watch & Subscribe Here

Have a product or service that would be great for our audience of vertical SaaS founders/operators/investors? Reply to this email or shoot us a note at ls@lukesophinos.com

I’m hosting a live workshop next week with the Co-Founder of an AI firm who has implemented agents 100+ times at A TON of comapnies. I think you’ll enjoy it…

If you’ve been curious about AI Agents and what they can do for your business, then this live workshop is for you.

We’re keeping it intimate and only allowing the first 25 founder sign-ups!
5 seats left.

Sign Up For Live Workshop

Alright, let’s get to it…

Embedded Payments Don’t Fail Because of Tech. They Fail Because We Treat Them Like a Feature.

Alright, let’s get to it.

Most vertical SaaS teams don’t choose to underperform in payments.

They do the “right” stuff: pick a processor, ship the integration, update onboarding docs, train support, announce it to customers… and then sit back waiting for adoption to roll in like it’s a product-led growth miracle.

And when it doesn’t, the story we tell ourselves is almost always external.

“Merchants don’t like change.”
“This vertical is unique.”
“Payments just takes time.”

Sometimes that’s true. But if you’ve been around vSaaS long enough, you start noticing a pattern that’s way more uncomfortable:

Embedded payments rarely fails because the technology doesn’t work. It fails because leadership launches it like a feature instead of building it like a business.

And those are two very different mindsets.

The Companies Doing Real Numbers in Payments Made a Decision (Not a Guess)

The vSaaS platforms generating 40–70%+ of total revenue from payments didn’t stumble into it. They didn’t “add payments” and accidentally wake up with a money-printing machine.

They made a deliberate call that payments would be a core business line with:

• a clear owner (not “shared across product + partnerships”)

• a real number they obsess over (attach rate, activation speed, GP $)

• incentives tied to outcomes

• executive visibility every single week

When payments is treated like a side quest, adoption stays optional. When payments is treated like a business line, the organization starts behaving differently—because it has to.

The Shift: From “How Do We Add Payments?” to “How Does Payments Change the Business Model?”

Low-performing teams ask: “How do we integrate payments?”

High-performing teams ask something more strategic: “If payments works, how does it change the economics and operating model of our SaaS?”

That question forces decisions you can’t dodge.

Product stops treating payments like a checkout screen and starts treating it like a merchant experience. Not just “can they run a card,” but “how fast can they get live,” “do they trust payouts,” “does reconciliation make their bookkeeper less angry,” “does this feel safer than what they’re doing today.”

Sales and Customer Success stop pitching it like an add-on and start treating it like a required part of the platform’s success. Not in a sleazy way—more in a “this is the modern way the software works” way. The best teams don’t hope merchants adopt. They build a motion where merchants naturally end up there because the path is clear, supported, and repeatedly reinforced.

And finance—this one matters—stops seeing payments as pass-through revenue. Because it isn’t. Payments is margin, risk, chargebacks, underwriting, pricing strategy, and (if you’re doing it right) a meaningful driver of LTV expansion.

You can’t run that like a feature.

Marketing Isn’t Optional. It’s the Adoption Engine.

Here’s where a lot of SaaS leaders get it twisted: they think payments adoption is a sales problem.

It usually isn’t.

Yes, sales helps. Yes, CS helps. But the strongest vSaaS payments motions are marketing-led and sales-supported, because what you’re really doing is changing behavior. You’re asking a business to re-route how money moves. That’s not a “new feature.” That’s trust.

So what actually drives adoption?

Repetition and clarity.

Merchants adopt when the value is so obvious they feel dumb not doing it. They adopt when the story shows up in-product, in lifecycle emails, in webinars, in case studies, and in conversations with your team—consistently.

And importantly: not “Now with payments!”

More like: “Want faster cash flow?”
“Want less admin work?”
“Want cleaner reconciliation?”
“Want to stop chasing invoices?”

You’re not selling payments. You’re selling outcomes.

Say whaaaaaaaat—marketing actually matters.

Incentives Tell You What the Company Really Believes

If payments revenue isn’t reflected in compensation plans, it’s not a priority. Full stop.

You can write all the strategy memos you want. But if AEs don’t get paid on attach, if CSMs don’t get measured on activation, if exec meetings never review payments KPIs, then payments is just a thing you integrated… not a business you’re building.

When incentives line up, the whole machine tightens:

Sales cares earlier.
CS cares longer.
Product cares about friction.
Leadership cares about the weekly number.

That’s when adoption stops being “nice to have” and becomes inevitable.

Rails Are Table Stakes. Operator Experience Is the Differentiator.

Most providers can give you processing rails. Cool. That’s the minimum.

What matters is whether the partner (and your internal team) actually understands the messy operator reality of vertical SaaS: onboarding flows, underwriting edge cases, support volume, disputes, pricing packaging, renewals, and the constant tension between growth and risk.

One example that’s interesting here is Xplor Pay—not because they have magical APIs, but because they are the only payments provider I’ve ever seen that actually operates vertical SaaS businesses.

They do both so they get it. Having built and scaled 20+ vertical SaaS platforms and have seen attach rates as high as 95% in some verticals. That kind of performance typically comes from getting the full system right: product + GTM + incentives + operations, not just “we embedded a payments button.”

The Takeaway

Embedded payments doesn’t succeed because you integrated it.

It succeeds because leadership decides to run it like a business line—complete with ownership, economics, marketing-driven adoption, and incentives that force focus.

If your payments attach rate is stuck right now, ask yourself the real question:

Are we dealing with a product friction problem… a value communication problem… or an incentives problem?

You’ve got this.

What do Org Charts look like in an AI world?

The old way of building vertical software was pretty linear:

Build a wedge product → add modules → sell seats → become the system of record

It wasn’t wrong. It was just… the best we could do.

The new way is different:

Software doesn’t just track the work.
It does the work.

That’s the shift LLMs enabled: automating tasks that traditional vertical SaaS could never touch. And the budget you’re competing for changes too, from IT to labor.

And once you’re replacing “1–2 FTEs worth of work” (or pricing like you are), the org chart changes.

So what are the roles that make sense in this new age? Org Charts will never look the same again...

A few ideas:

1) Industry Expert
The person who knows the ugly edge cases, compliance gotchas, and what users actually do at 6:30am. No operator = no truth.

2) Workflow Architect
Old product shipped screens. Took them weeks/months.
New product shows full UI prototypes and screens to customers immedietly. Explains what the agent does, what the human is tasked with. Where the ROI ends up.

3) Ontology / Data / Agent Knowledge Expert
If your AI is “doing work,” it needs a real-world map: entities, states, definitions of “done,” and what “correct” means in this industry.

4) Applied AI Engineer
This is the builder turning “cool demo” into production ready software. They should be in every meeting with the Workflow Architect and the customer. Engineers hiding in ivory towers and behind PM's is OVER.

6) Trust Lead (Security / Compliance / QA)
You need someone accountable for guardrails, auditability, and what happens when the model is wrong.

7) AI GTM AE
This is a sales person that uses all sorts of AI tools and Agents to get prospects into meetings. Lazily sifting through the CRM and sending a few handwritten emails is DEAD.

8) Adoption Owner (CS that ships outcomes)
Old CS: QBRs + renewals.
New CS: instruments usage, drives behavior change, automates ROI reporting.

If you’re building vertical software today…
The question 2 years ago was “should we add AI?”

Now it should be...
Is our Org Chart actually designed for a radically different business?

It’s EPIC and I just had to share it with you all…

But before we dive into the report,
I’m hosting a live workshop next week with two killers and think you will really enjoy it….

Building & Implementing AI Agents — Deploy AI agents that actually work

If you’ve been curious about AI Agents and what they can do for your business, then this live workshop is for you. We are going to break down how to deploy agents that ACTUALLY drive measurable outcomes, ie growing revenue, decreasing costs, maintaining compliance.

We’re keeping it intimate and only allowing the first 25 founder sign-ups! Click below to sign-up.

Sign Up For Live Workshop

2025 was the year Vertical AI moved from hypothesis to reality. After a 2024 defined by AI infrastructure buildout and foundation model wars, capital flowed decisively toward industry-specific use cases. We spent months dissecting every software and AI venture deal of $1M+ across the US and Canada — 4,395 financings and 254 exits, analyzed by vertical, stage, investor, geography, and more — to understand exactly what happened, and what it means for 2026.

Here are a few of our most interesting findings, in brief:

The Year in Numbers

The US and Canadian VC market produced 4,395 software and AI financings totaling $186B in 2025. Vertical startups captured 53% of deal volume and 30% of capital deployed—though that capital gap is almost entirely a function of a small handful of infrastructure mega-rounds (OpenAI’s $40B, Anthropic’s $13B, and a few others). Strip out just the 12 deals sized $1B+ and vertical takes 51% of capital.

Deal volume was relatively even across quarters—Q1 and Q2 combined for 52% of the year’s deals—but capital deployment was lumpy. Q1’s $60.5B was inflated by OpenAI’s Series F alone; remove it and the quarter drops to roughly $20B, making the real quarterly trajectory far more balanced. Vertical’s share of capital climbed steadily: 19% in Q1, 34% in Q2, 32% in Q3, and 42% in Q4.

On the exit side, 2025 delivered $132.7B across 158 vertical transactions—plus another $102B in non-vertical exits. M&A dominated by count, but the story is more nuanced than that, as we’ll get to below. Healthcare led exit volume with 43 transactions, and the IPO window cracked open with 18 vertical offerings after a two-year drought.

Our full report covers vertical-by-vertical breakdowns, stage dynamics, investor league tables, geographic distribution, profiles of fifteen notable deals, and our 2026 outlook. What follows are five findings from the data that we think reframe the conventional wisdom and set the stage for another banner year in Vertical AI.

1. The Remarkable Consistency of Mega-Deals

Most people assume late-stage venture is the volatile tier—big rounds, big markdowns, boom-bust cycles. In 2025, the opposite was true.

The $100M+ tier of vertical financings was the most stable bucket all year. Since Q1, we saw 55-57 deals every quarter. Meanwhile, the $1–5M tier of deals — pre-seed and early seed—dropped from 503 deals in Q1 to 362 in Q4, a 28% decline. The bottom of the market pulled back far more than the top.

Who’s sustaining that stability at the top? The investor data tells the story. The mega-round tier ($100M+) was anchored by a small group of multi-deal leads: NEA (10 leads), Kleiner Perkins (7), Oak HC/FT (5), B Capital (5), DST Global (4), and Ribbit (3). Their average deal sizes clustered tightly in the $110–180M range.

Beyond the frequent flyers, the mega-deal lead base broadened: BlackRock co-led Applied Intuition’s $600M round, Franklin Templeton led Plaid’s $575M, Qatar Investment Authority and Mubadala showed up in multiple rounds. Asset managers in pre-IPO rounds is reminiscent of 2021 — the uptick of sovereign wealth involvement is perhaps more notable, especially in light of their parallel growing interest in backing US VC funds as LPs.

2. A&D Momentum: Conviction in Chaos

When people think “hot vertical AI categories,” they think healthcare and fintech — and of course, they’d be right (especially regarding the former). But the 2025 data tells us that another category may be dominating the hype cycle.

Aerospace & Defense led all verticals in both capital intensity per employee ($476K per FTE) and median valuation step-up (2.15x). It beat Manufacturing ($421K, 2.08x), which got most of the headlines thanks to Project Prometheus ($6.2B) on the funding side and Ansys ($35B) on the exit side. It beat Financial Services, which despite attracting the most aggregate capital posted a modest 1.80x step-up (turns out bigger rounds dilute per-round markups even when total value creation is strong).

The contrast with Education is particularly stark. Education had the lowest capital intensity ($200K per FTE), the lowest capital velocity ($1.6M raised per year since founding), and the lowest capital efficiency ($342K raised per employee)—bottom of the table on every capital metric. And yet Education posted the highest early-stage step-up (2.5x in the $5–15M tier). Despite great emerging stories in SchoolAI and MagicSchool, the category isn’t yet a proven absorber of capital at scale. A&D is the opposite: capital-hungry and valuation-rich, despite its unproven track record as a generator of historical VC liquidity.

Forterra’s $238M Series C—profiled in our full report—is a case in point. Unlike some other A&D high-flyers, they have production contracts and program-of-record revenue, not R&D grants. Backed by an interesting investor syndicate (Salesforce Ventures, Franklin Templeton, Hanwha, NightDragon) that reflects the reanimation of crossover investing. Beyond any one deal—and equally relevant to VC and LP funding—the intersection of government and technology a is dynamic we’re watching closely heading into 2026.

3. CA: America’s Least & Most Vertical State

Amongst all US states in 2025, California posted the lowest ratio of vertical deals, 41% — yet it was also home to the highest absolute number, at 667 deals.

This makes CA simultaneously the least and most vertically oriented in the country. Non-coastal standouts—Arizona (71%), Ohio (70%), and Tennessee (67%)—posted vertical shares that make California look like it’s playing a different game. While New York, at 53%, tracked near the national mean in terms of verticality, it was the clear #2—home to 406 vertical deals in 2025, with signs of the most momentum.

The explanation, in CA’s case, is straightforward: dominant concentration of capital-hungry horizontal AI infrastructure, foundation models, and dev tooling pulls the denominator up. And while it makes sense that company formation is strongest in industry-proximate geographies—contractors in Arizona, manufacturers are in Ohio, healthcare operators in Tennessee.

The regional data adds another layer. The Midwest leads in vertical proclivity at 59.6%, followed by a tight cluster in the mid-fifties: Mid-Atlantic, Mountain West, and Southeast. The West trails at 41.3%. And Canada’s lower vertical share—somewhat surprising given Toronto’s legacy as home of Constellation Software—appears driven by concentration in a single metro, with much of its vertical reach involving investment in US-based companies.

San Francisco and New York remain the twin dominant vertical hubs by volume and deal quality. And while the data show that non-coastal metros embrace verticality, they have two big hills to climb: (1) bridge the massive gulf of general startup gravity; and (2) keeping pace on AI-nativity in this dynamic new environment.

4. Manufacturing is the Momentum Leader

We examined the share of 2025-funded companies in each vertical that were founded in 2022 or later—the post-LLM cohort. Manufacturing & Industrial led in “AI-native” share at 57%. Not healthcare. Not legal. Manufacturing.

More than half of all manufacturing AI companies funded last year didn’t exist before ChatGPT. The cross-vertical baseline is 46%, meaning manufacturing’s new-company share runs 11 points above average. Public Sector (56%) and Legal (53%) also exceeded the mean, but manufacturing’s lead reframes the narrative around the dominance of white-collar verticals in AI-native opportunity.

Deal activity data supports the formation data. Manufacturing deal count rose 41% from Q1 to Q4, the steepest acceleration of any vertical. The category produced the year’s second-largest financing—a $6.2B round we profile among the full report’s fifteen deal highlights.

At the same time, it’s important to remember that as AI and robotics merge into “Physical AI,” core differences in business model can be blurred. Much of this activity in Manufacturing — at least from a capital standpoint — is going into hardware-dominant, high marginal-cost businesses. At $421K in capital intensity per employee, Manufacturing & Industrials companies are raising more to meet the physical-world infrastructure requirements of bringing AI to factory floors, supply chains, and industrial operations — and they are spending more too.

A quick word from our sponsor…

Verticals is made possible by Parafin—embedded capital for your merchants.

If you’re building a vertical platform, embedded finance isn’t just “another product.” It’s one of the most consistent ways we’ve seen to drive retention and unlock a meaningful new revenue layer (without trying to squeeze blood from a seat-based pricing stone).

They’ve powered financing inside platforms at massive scale (yes, the big guys) and the core idea is simple: help your customers access capital inside the workflow they already live in.

5. Liquidity, Liquidity, Liquidity

In 2025, buyout and LBO transactions accounted for 20% of vertical exits by count—but generated about as much liquidity as M&A. PE firms view vertical software as ideal acquisition targets: recurring revenue, defensible market positions, and operational leverage from AI integration. The majority of PE-backed software acquisitions in 2025 targeted vertical companies. Silver Lake, EQT, and TPG were among the most active acquirers.

Meanwhile, IPO activity last year was meager — but perhaps a start. 2025 produced 9 vertical IPOs (of 15 total) generating $4.3B in aggregate value (just under 40% of all IPO value). Something after a two-year drought, but thin relative to the backlog of late-stage companies that raised $200M+ rounds over the last 12-24 months. Expect more tough go/no-go listing decisions in 2026–2027.

The most revealing exit metric may be the simplest one: the median vertical exit in 2025 was approximately $500M. For the many seed and Series A funds that have scaled up AUM since 2021, driving superlative returns via such median outcomes demands discipline on entry price that seems quite rare these days. At a $10-20M post-money seed, a $500M exit represents 25-50x gross valuation appreciation—exceptional. At a $50M+ post-money “seed” (increasingly common in consensus Vertical AI), the same exit is 7-8x gross. Still great, but tough math for a >$100M AUM fund.

The top 10 exits by value in 2025 accounted for over 60% of total vertical exit value—a reminder that no category escapes power-law dynamics. Below the mega-deals, PE remains a leading exit path, IPOs remain selective — the “middle market” of vertical exits ($250–1B EV) is arguably the healthiest and most under-appreciated segment of the liquidity environment, at least in VC land.

What’s in the Full Report

The findings above are drawn from roughly a quarter of the complete Vertical Report. The full publication includes:

• Vertical-by-vertical analysis of 18 sectors: deal count, capital deployed, stage mix, median deal sizes, and quarterly momentum for each—from Healthcare (597 deals, $11.8B) down to Energy & Climate (60 deals, $900M).

• Stage dynamics deep-dives: deal size distributions by vertical, capital intensity, valuation step-ups, capital velocity, and capital efficiency metrics.

• Profiles of 15 notable deals, from a $190M public-sector round to a $6.2B industrial mega-financing—covering step-ups, investor syndicates, founder backgrounds, and what each deal reveals about its vertical.

• The full exit landscape, including >150 deals worth $130B+ broken down by type, quarter, vertical, and size—plus the IPO pipeline and PE acquisition patterns.

• Investor league tables, from the most active seed leads to the dominant growth-stage firms—plus corporate venture and SWF trends.

• Complete geographic analysis by state and region.

• Our 2026 outlook on four structural dynamics: the seed-to-A graduation test, deep tech’s maturity cycle, shifting liquidity winds, and the coming fight for AI defensibility.

The report is too long for email.
So if you want to continue go check the rest of it out at Nic’s newsletter here.

Have a product or service that would be great for our audience of vertical SaaS founders/operators/investors? Reply to this email or shoot us a note at ls@lukesophinos.com

Today’s newsletter is sponsored by Xplor Pay. If you’re attending TRANSACT 2026 in Atlanta, stop by Booth #717 to learn how you can unlock new revenue streams, scale faster, and streamline the payment experience.

They’re hosting an espresso bar in their booth on Thursday, March 19th, starting at 9 a.m. Stop by if you’re at the show and receive a surprise gift! Learn about PayFac as a Service with Xplor Pay.

Alright, let’s get to it…

Prepared acquired for $640M

Prepared was the wedge.
Axon is the system of record.
That’s the story.

I wrote about Prepared back in Linear #137 (“Prepared: VerticalAI For 911 Dispatch”) because they were a perfect example of how you actually win in a vertical where “selling software” is the easy part and “changing behavior” is the hard part.

And now they’ve been acquired by Axon in a $640M acquisition. Here’s the non-obvious point that matters for every vSaaS founder/operator reading this:

This wasn’t “a 911 software company acquired by another 911 software company.”

This was a wedge product getting absorbed by the system of record / platform incumbent. Prepared didn’t become the big dog. They became the thing the big dog couldn’t ignore.

What Prepared actually was: the “first 120 seconds” wedge

Prepared’s wedge is brutally simple and strategically devastating: it lives inside the earliest moments of an incident — the moment when chaos becomes “what we think is happening.”

Axon frames it as “owning the first 120 seconds.” In that window, Prepared is translating, transcribing, summarizing, flagging urgency, and turning raw signals into usable intelligence.

That matters because the first 120 seconds are upstream of everything.

Upstream of dispatch. Upstream of response. Upstream of evidence. Upstream of reporting. Upstream of liability.

And Prepared had real distribution. The acquisition press release calls out 1,000+ agencies across 49 states and nearly 100 million people protected.

So Prepared wasn’t just “a feature.” It was becoming the interface for the highest-stakes workflow in the public safety category.

What Axon is: the system of record (and the big legacy dog with all the other stuff)

Now zoom out.

Axon isn’t a one-product SaaS company. They’re the platform incumbent in public safety tech — the company most people associate with Tasers and body cams, plus the software stack that sits behind those devices.

This is what systems of record do in real life: they don’t just store data. They become the procurement standard, the compliance anchor, the training layer, the vendor your agency can’t rip out without a multi-year migration and a political fire drill.

So when a wedge product like Prepared starts to own the front door of incident response, the system of record has a choice:

1. Let the wedge become the UI (and eventually the control point), or

2. Buy it and make it part of the platform narrative.

Axon chose option #2. The press release makes it explicit: Prepared strengthens Axon’s strategy to connect public safety “from call to closure.”

That phrase is doing a lot of work.

Because “call to closure” is another way of saying: “we’re the operating system.”

Here’s the deeper trend: every legacy SoR is panic buying the AI agent layer

This is where the Prepared/Axon deal stops being “public safety news” and becomes “software industry inevitability.”

The agent layer is becoming the new UI. And the new UI is where the power lives.

If you’re the system of record, that is terrifying.

So you’re watching systems of record across the enterprise do the same thing Axon just did:

They’re acquiring the wedge / agent companies so they control the front door.

1) Procore (construction SoR) → acquired Datagrid (agentic AI wedge)

This is the most on-the-nose comp to Axon/Prepared.

Procore is the construction management system of record for a massive chunk of the industry. Then they went out and bought Datagrid, explicitly positioning it as a way to eliminate data silos and automate workflows like submittals + RFIs across platforms (i.e., “initiate actions,” not just store data).

Procore even calls out the agentic angle: Datagrid adds “advanced reasoning,” “deep search,” and the ability to unify intelligence across third-party systems like ERP and cloud storage — classic “agent layer sits across your stack” behavior.

2) ServiceTitan (trades SoR) → acquiring Conduit Tech (AI-enabled field/sales wedge)

ServiceTitan is the system of record for home services ops (dispatch, CRM-ish workflows, invoicing, etc.). Conduit is a wedge closer to the field that makes it easier to scan a home (LiDAR), produce designs, and generate proposals fast—which directly impacts conversion and revenue per lead.

ServiceTitan’s framing is telling: they’re integrating Conduit into their suite of “AI automation products,” and they’re buying leverage at the moment value is created (assessment → proposal → close), not just the back office.

3) Guidewire (P&C insurance SoR) → acquiring ProNavigator (AI knowledge agent layer)

Guidewire is a core system of record in P&C insurance (policy, claims, underwriting workflows). They signed a deal to acquire ProNavigator, which is an AI-powered knowledge management platform that delivers “context-aware” best-practice answers inside the workflow.

That’s the wedge: not replacing the SoR, but becoming the thing users ask first while working in the SoR.

This is the same movie over and over:

• A wedge product shows up and says: “We’ll take the messy unstructured inputs and do the work.”

• Customers love it because it actually reduces labor.

• The system of record realizes the wedge is becoming the interface.

• The system of record buys the wedge.

Prepared just did that in a vertical where the outcome is measured in minutes and lives, not clicks and conversion rate.

Why the “SaaSpocalypse” narrative makes this intensify (not slow down)

The whole “SaaSpocalypse” thing gets memed to death, but the direction is right: buyers are tired of paying for software that just creates more work.

They want systems of action. They want “less labor.”

And that’s why the agent layer is so dangerous to the system of record. As NEA puts it, we’re shifting from “system of record” toward “system of actions,” where AI agents do more of the remaining labor-heavy workflow.

If that’s true (and it is), then the SoR can’t just integrate the agent layer.

It has to own it.

Otherwise, the SoR becomes a backend database with legacy switching costs… until the day the UI moves, users move, budgets move, and procurement follows.

The punchline for vSaaS builders

If you’re building in Vertical AI right now, the question isn’t “can I build an agent?”

The question is: can I wedge into the workflow early enough that I become strategically unavoidable to the system of record?

Prepared did that by taking the first 120 seconds — the moment when everything downstream gets shaped.

Axon did what big systems of record do: it bought the control point before someone else owned it.

And the rest of the Vertical SaaS market is following…

Vertical Software/AI Summit 2026 is going to be BIG

There are a few other folks out there throwing Vertical SaaS conferences. I think this is great, the more the merrier. It’s about time we had our time.

For those trying to pitch VC’s on Sand Hill Road from 2010-2017, and getting laughed out of the room in a pre-scale Toast, Service Titan world you can relate….

The only thing I’ll say is

A) We are the only conference by Vertical SaaS founders FOR Vertical SaaS founders

and

B) I’m going all out on this years conference and expecting 400-500 vSaaS founders + the entire investment community to be there.

I have ~5 unicorn founders + one billionare already confirmed to speak 🤯

It’s going to be epic. So tap the link below and be ready for an epic few days in Q4…

Sign Up To Be Notified When Tickets Drop

PS - I will give free tickets to a bunch of folks that opt in above :-))

Charge An Implementation Deposit

This one is stupid simple.

And it massively moved the needle for me at my last company.

Here’s the idea:

You charge a refundable implementation deposit equal to ~5% of the annual contract value.

You agree on a specific milestone that represents “real adoption.”

If the customer does what they said they’d do (give you data on time, show up to training, assign an internal champion, etc.), they get the deposit back.

And to be crystal clear: you should basically always give it back. This is not a revenue driver. This is a commitment device.

It’s you saying: “We’re going to implement you successfully. You’re going to participate.”

Why it works (and why it’s not just a money grab)

In vertical SaaS, implementations fail A LOT.

Especially with Mid-Market + Enterprise customers.

It’s typically not product, not integrations, not “edge cases.”

It fails because the customer doesn’t do their part.

They don’t get you the roster/export/file.

They don’t schedule training.

They don’t enforce the new software/AI.

They ghost… and then 90 days later you’re staring at a churn risk that was created on Day 7.

The deposit fixes the psychology.

The moment a customer has real money on the line (even a small amount relative to the contract), implementation stops being “that thing we’ll get to” and becomes “a project we need to finish.”

And that changes everything.

The exact structure I like

Deposit amount: 5% of annual contract value.

Collected: upfront, at signature (or before kickoff).

Refund mechanism: refund it when the customer hits the agreed milestone.

Important nuance: the refund can be a literal refund or a credit on the next invoice—depending on how your billing/procurement reality works. But make it clean and explicit.

And again: you are not trying to “catch” customers and keep their money.

You are trying to create a forcing function that makes adoption inevitable.

Pick a milestone that signals “you’re live,” not “we had meetings”

If your milestone is “first kickoff call completed,” congrats, you’ve invented a useless deposit.

You want a milestone that requires customer behavior.

A few examples that work well:

Milestone option A (best for workflow products):
Customer is live with X users and completes Y workflows in production (not sandbox) over a 2-week period.

Milestone option B (best for data-heavy products):
Data import complete + training complete + first “real” report/transaction submitted from the system.

Milestone option C (best for multi-site / enterprise):
First site live with adoption KPI met (then you expand to the rest).

Notice the pattern: the milestone is observable, and ideally quantitative, not vibes.

The customer commitments need to be written down

This is the part people skip.

If you’re going to ask for a deposit tied to a milestone, you need the customer to agree to what they will do.

I’d define “Customer Responsibilities” in plain English, like:

• Name an executive sponsor and an internal project owner

• Provide required data by a specific date

• Ensure attendance at training (with names listed)

• Confirm internal rollout date

• Respond to implementation questions within X business days

• Agree to a real definition of what success looks like!

If they don’t do those things, implementation doesn’t move.

That’s not you being annoying. That’s reality.

What changed for me when we did this

The most surprising impact wasn’t cash flow (again, you give it back).

The impact was operational:

• Implementations stopped dragging forever

• Customers showed up prepared

• “We’re too busy” became “we’ll get it to you by Thursday”

• Internal champions suddenly had teeth because they could say, “We already paid the deposit—let’s finish this.”

It also quietly improved retention, because successful implementations are the first domino in long-term customer health.

If you’ve ever watched a customer churn 6 months in and thought, “Yeah… this was doomed from the beginning,” you know exactly what I’m talking about.

How to position it so customers don’t freak out

Call it what it is:

A refundable implementation deposit that ensures both sides are committed to a successful launch.

Here’s the script:

“We charge a 5% implementation deposit upfront. It’s fully refundable once we hit the go-live milestone we agree on together. The reason we do this is simple: implementations only work when both sides prioritize the project. This keeps us aligned, and it keeps your team from losing momentum.”

If you say it calmly, like it’s standard (because it should be), serious buyers respect it.

And if someone loses their mind over it? That’s usually a signal you were about to get a slow, painful, low-adoption customer anyway.

The two common pitfalls

Pitfall #1: treating it like revenue.
Don’t. If your team starts celebrating kept deposits, you’ve turned a customer-success mechanism into a weird punishment system. You’ll lose trust fast.

Pitfall #2: picking the wrong milestone.
If the milestone is too easy, it won’t drive behavior. If it’s too hard or too subjective, procurement will block it and customers will argue.

Make it measurable. Make it fair. Make it tied to adoption.

One tactical detail (that saves headaches): keep the refund process automatic

If the customer earns the refund and then has to chase you for 6 weeks, you just created unnecessary resentment.

Operationalize it:

When milestone is achieved → implementation lead marks complete → finance triggers refund/credit within X days.

Make it boring.

Make it consistent.

Charging an implementation deposit is one of those operator moves that doesn’t show up in product screenshots, but changes the trajectory of accounts.

And if you can make launches predictable, everything downstream gets easier: retention, expansion, referrals, your team’s sanity.

Have a product or service that would be great for our audience of vertical SaaS founders/operators/investors? Reply to this email or shoot us a note at ls@lukesophinos.com

Since launching Verticals late last year, we’ve had the honor of hosting ~20 founders, operators, and VCs who have collectively built billions in vertical enterprise value.

And honestly? It’s been one of the most fun quarters of content I’ve ever been a part of.

Not because it’s flashy.

Because it’s real.

It’s operators talking about what worked, what didn’t, and what they’d do differently if they had to build the same business again—except now the whole world is screaming “AI!” while you’re just trying to ship product, close deals, and not blow up retention.

If you’ve been watching on YouTube or listening on Spotify, you’ve already caught a bunch of these moments in real time. This post is a look back at the best ones—and the patterns that keep showing up.

First, a quick word from our sponsor

Verticals is made possible by Parafin—embedded capital for your merchants.

If you’re building a vertical platform, embedded finance isn’t just “another product.” It’s one of the most consistent ways we’ve seen to drive retention and unlock a meaningful new revenue layer (without trying to squeeze blood from a seat-based pricing stone).

They’ve powered financing inside platforms at massive scale (yes, the big guys) and the core idea is simple: help your customers access capital inside the workflow they already live in.

The big meta-lesson: vertical AI isn’t “a category,” it’s a forcing function

One thing keeps coming up across episodes:

AI isn’t killing vertical SaaS.

It’s exposing vertical SaaS.

If your product is a thin point solution with no distribution advantage, no workflow ownership, and no credible path to becoming a true platform… it’s going to feel rough out there.

If you’re building something that sits on top of the system of record, owns the workflow, and compounds into payments / capital / services / network effects?

You’re going to look back and realize this was the best possible time to be building.

That’s why we spent an entire episode debating the “SaaSpocalypse” framing. The bears say SaaS is dead. We disagree. The market is expanding—but the winners will look different, and “distribution” is back on its throne.

Episode 1: Ilir Sela, Slice — “Independent, not alone” is the whole game

Ilir basically gave us the thesis for the show in Episode 1.

Slice is the “reverse franchise” model in its cleanest form: give independent operators the shared stack and economies of scale of a Domino’s without forcing brand uniformity.

The moment I can’t shake is his voice ordering point.

Slice’s AI voice ordering hit ~90% transaction completion, but humans were still at 98%. That 8-point gap doesn’t sound like much until you realize it’s the difference between a profitable shop and a failed shop in low-margin land. So until AI clears the bar, humans stay in the loop.

That’s the kind of cold pragmatism more “vertical AI” decks need.

Also: Ilir’s pricing and ACV progression is a masterclass. Slice started with a performance-based model—$2 per order—then layered into payments and additional products over time.

If you’re building in a fragmented SMB market and you’re not thinking “reverse franchise / business-in-a-box / outcome-aligned pricing”… what are we doing?

Episode 2: Jeremy Yamaguchi, Cabana — AI roll-ups are three businesses in a trench coat

AI-first roll-ups are one of those ideas that sounds simple on Twitter and is brutally complicated in real life.

Jeremy put it plainly: you’re running three businesses at once—an operating company, a tech company, and a PE shop.

That’s why this model is so misunderstood. People argue “VC vs PE” like it’s ideological. In reality, it’s operational: can you source deals, integrate teams, standardize workflows, and still ship product fast enough that your “AI margin uplift” isn’t a powerpoint fantasy?

The upside is clear though: if you own the P&L, you can force adoption in verticals that might otherwise resist software change for a decade.

Episode 5: Todd Saunders, Broadlume — community can be a moat (if you earn it)

Todd’s story is one of the best reminders that “boring vertical” is often where the real enterprise value hides.

Broadlume built a real platform in flooring by leaning into community, content, and first-party data—and doing it with a level of vertical intimacy that generic SaaS teams just don’t have the patience for.

And importantly: this wasn’t a story about avoiding risk. Todd talked openly about making painful calls and enduring major pivots to build something durable.

Broadlume ultimately became valuable enough that Cyncly entered into an agreement to acquire it, highlighting the strategic value of owning the digital marketing + visualization + business management stack in a vertical with thousands of dealers and manufacturers.

If you’re early, the question is: are you building “software,” or are you building an institution in your vertical?

Episode 8: Sam Youssef, Valsoft — “small TAM” is often where the compounding machine lives

Sam runs Valsoft—130+ acquisitions and ~$750M+ revenue—and his view on market selection is beautifully contrarian.

The bigger the vertical, the less interesting it often is.

Why? Big markets attract irrational competition and giant venture checks. Smaller verticals tend to stay rational—pricing, competition, budgets, expectations.

And Sam’s also refreshingly honest about the reality that even great acquirers don’t bat 1.000. A meaningful chunk of acquisitions underperform.

That’s not a reason to avoid vertical. It’s a reminder to respect it.

Vertical is not easy mode.

Vertical is a different sport.

Episode 10: ServiceTitan fintech — do customer value first, take rate second

One recurring theme across the best platforms: the fintech layer is where your model can go from “good SaaS business” to “category-defining company.”

But the sequencing matters.

The best operators don’t wake up and say “let’s add payments so we can juice ARPU.”

They say: let’s obsess over customer pain, solve the workflow, earn trust, and then monetize money movement because it’s the natural extension of the system.

In other words: value first, take rate second.

If you try to reverse that, you’ll feel it in churn and NPS long before you see it in attach rate.

Episode 13: Scott Wolf, Levelset — SMB is often a disguise for enterprise economics

Scott’s episode is one every vertical founder should watch if you’ve ever heard the sentence: “Your customers are too small.”

Levelset sold to subcontractors, sure—but Scott made the nuance crystal clear:

At exit, their SMB ACV was roughly $4.5k–$5k, and they had enterprise supplier contracts that were seven figures.

Say whaaaaaaaat.

That’s the play: wedge with the persona that gives you distribution, then monetize deeper in the value chain where budgets and leverage live.

Also, his M&A framing is evergreen: “Great companies are bought, they’re not sold.”

Build something strategically unavoidable and the market comes to you.

Episode 14: Nate Baker, Qualia / Fractal — TAM isn’t fixed, it expands as you earn workflow

Nate challenged one of the most persistent objections in vertical: “the TAM is too small.”

He’s built Qualia at real scale—500K+ professionals served and ~600 employees—and he’s helped start 150+ vertical SaaS companies through Fractal.

His point is the one I wish more founders internalized earlier:

In vertical, TAM expands when you create new layers of value.

You start with a wedge, you earn the workflow, you become the system of record, then you layer payments, capital, services, and ecosystems.

That’s how the “small market” becomes a monopoly-like outcome over time—Nate even points out that vertical SaaS can reach 50–70% market share in a category.

And yes, he had a nuclear quote: every business that isn’t reevaluating its product offering right now is going to be cooked.

He’s not wrong.

Episode 15: EvolutionIQ — the exit story everyone forgets: it often starts tiny

One of my favorite reminders from the show is that many great companies start with what looks like nothing.

One design partner.

A tiny first contract.

A painfully manual workflow.

Then the wedge hardens, the platform emerges, and the market wakes up.

EvolutionIQ is now part of CCC Intelligent Solutions—CCC completed the acquisition and explicitly framed it as expanding AI capabilities like medical summarization and “next best action” for claims management.

Founders: don’t let today’s “revenue bar” narratives trick you into thinking you need perfection at inception.

You need a wedge that earns trust.

Episode 17: Darren Fike, Santé — AI isn’t just product, it’s a GTM weapon

Darren had one of the most tactical insights of the entire season:

Use AI to collapse switching costs.

If your prospects have been on the same system for 10+ years, the biggest competitor is not the incumbent. It’s inertia.

Santé uses AI to pull prospect data and show the platform pre-populated during the demo, reducing the psychological distance between “this looks cool” and “we can use this tomorrow.”

And yes, the episode write-up notes they closed 30 deals in a single month with zero feature requests—which is one of the cleanest “PMF signals” you’ll ever see.

If you’re building vertical AI and you’re only thinking about “AI inside the product,” you’re leaving a weapon on the table.

Episode 18: Nick Tippmann, TipTop VC — distribution is back, and it never left

In the middle of the “SaaS is dead” discourse, Nick came on and brought the conversation back to earth.

Yes, public markets can evaporate hundreds of billions in weeks.

Yes, AI changes pricing models.

But the question isn’t “will software survive?”

It’s “who wins?”

And the winners tend to do the boring stuff religiously: instrument RevOps early, sell revenue-connected wedges, build trust in the industry, and get distribution right.

You can have the best model in the world.

If you don’t have distribution, it doesn’t matter.

One final thought

On behalf of the whole VERTICALS crew: thank you for reading, watching, listening, sharing, and DM’ing us your ideas.

At least in my career, there’s been no more exciting time to be part of vertical software and vertical AI.

New episodes drop every Wednesday.

If you haven’t yet, please do us a favor and subscribe to Verticals here:

• YouTube: Verticals channel

• Spotify: Verticals on Spotify

And if you’re an operator building something new in Vertical AI, send it our way.
We love this stuff.

You’ve got this!

Have a product or service that would be great for our audience of vertical SaaS founders/operators/investors? Reply to this email or shoot us a note at ls@lukesophinos.com

Today’s newsletter is sponsored by Xplor Pay. If you’re attending TRANSACT 2026 in Atlanta, stop by Booth #717 to learn how you can unlock new revenue streams, scale faster, and streamline the payment experience.

They’re hosting an espresso bar in their booth on Thursday, March 19th, starting at 9 a.m. Stop by if you’re at the show and receive a surprise gift! Learn about PayFac as a Service with Xplor Pay.

Alright, let’s get to it…

Agents vs. Systems of Record:
What Will Happen?

Most companies have a “systems of record are holding our data hostage” problem.

Your ERP / EHR / LMS / ancient desktop app has the data BUT...
-no API
-exports are janky
-the only person who knows how it works is “Steve in accounting”
-every “integration project” turns into a 6-month archaeology dig

Agents can technically solve this.

They can:
1. Open the system (yes, the UI)
2. Navigate like a human would
3. Pull the fields you need
4. Reconcile the mess (“Acct #” vs “AccountID” vs “Customer Code”)
push clean data into the real workflow (CRM, warehouse, dashboard)

This is bigger than “integration pain.” It’s a strategic shift in what it means to be a system of record… and what it means to have a moat.

Part 1:
What this means for systems of record (SoRs)

Historically, systems of record won because they became the “source of truth” and then (intentionally or not) held the data hostage.

Not always maliciously. Often it’s just reality:

• The vendor never built the API surface area.

• The data model is too messy to safely expose.

• The product team optimized for “get the job done,” not “enable the ecosystem.”

• The business logic lives in UI flows, stored procedures, and tribal knowledge.

So the SoR becomes the place where work happens and also the place where data gets trapped.

The agentic wedge: “UI as the integration layer”

Agents change the rules because they can do what humans do:

1. Open the system (yes, the UI)

2. Navigate like a human

3. Pull the fields you need

4. Reconcile the mess (“Acct #” vs “AccountID” vs “Customer Code”)

5. Push clean data into the real workflow (CRM, warehouse, dashboards)

This is the same conceptual promise RPA has had for years—bridging systems when APIs aren’t available—but agents bring a more flexible “understand + adapt” layer on top of it. Oracle even calls this out directly: RPA “bridges systems that lack APIs” to keep data flowing end-to-end.

And UiPath makes the same point from a different angle: UI-level integration can be a low-risk way to extract and move data when legacy systems don’t expose modern interfaces.

So what happens to the SoR?

A bunch of SoRs are about to learn a hard lesson:

If your primary “defensibility” is data gravity + integration friction, agents are about to sandblast that friction down.

Because the buyer doesn’t actually love your ERP screen from 2009.

If an agent can reliably read + write through the UI, you’ve just lost one of the most underappreciated lock-in mechanisms in enterprise software.

Part 2: What this means for moats (and all the companies built on top of them)

Let’s name the moat that’s quietly powered a lot of enterprise software outcomes:

The “hostage data” moat

• “We can’t switch because all our history is in there.”

• “We can’t switch because integrations took forever.”

• “We can’t switch because the workflows are custom.”

• “We can’t switch because no one knows what will break.”

That moat has been real. It’s also been… kind of gross.

Agents don’t eliminate switching costs overnight, but they attack the most expensive part: integration and migration effort.

If a system has no API, historically you had two options:

• Pay for a brittle, expensive integration project

• Give up and keep doing manual work

Now there’s a third:

• Let an agent operate the UI and produce structured outputs + audit logs

This is why the “unlock isn’t the model.”

The unlock is the combo of:

• Goal-based execution (agents can actually do the thing)

• A thin layer of governance (permissions + approvals)

• An audit trail (so you don’t get fired)

And governance is not optional, by the way. Gartner is basically screaming this between the lines.

Gartner predicts at least 15% of day-to-day work decisions will be made autonomously through agentic AI by 2028 (up from 0% in 2024) and 33% of enterprise software applications will include agentic AI by 2028.

Also: Gartner predicts over 40% of agentic AI projects will be canceled by end of 2027 due to escalating costs, unclear value, or inadequate risk controls. Translation: if you can’t prove reliability + controls, you’re toast.

“But moats are dead!”

Moats aren’t dead.

Lazy moats are dying.

If your moat is:

• “We’re the only place the data lives”

• “Integrations are hard so customers stay”

• “No one wants to migrate”

• “Our UI is the workflow”

…agents are coming for you.

If your moat is:

• Deep workflow ownership

• Compliance + certifications + trust

• Distribution into a vertical

• Embedded payments / financial rails

• High-frequency operational loops

• Network effects (real ones, not vibes)

…you’re still in a good spot.

The winner moats shift from data custody to decision velocity + workflow control.

Part 3: The new battleground: “System of record” vs “System of action”

If agents can pull data out of any UI, then the real product advantage moves up the stack:

• Who owns the action?

• Who owns the approval chain?

• Who owns the exception handling?

• Who owns the audit trail?

• Who owns the user’s trust?

In other words:

The agent layer becomes the new orchestration plane

And that’s where a ton of new companies will get built.

Not “ChatGPT but for X.”

More like:

• “We turn your legacy finance system into an API by operating the UI safely”

• “We reconcile IDs and write back with approvals”

• “We provide agent observability, rollback, and compliance evidence”

• “We own the decision loop, not just the database”

This is also why the old RPA vs API framing is getting updated.

RPA works at the UI layer; APIs work at the backend. Agents can blend both—and pick dynamically.

What AI agent founders should build (if you want to win)

Here’s the playbook I’d run if I were building an agent company attacking SoR-hostage problems:

1) Pick one “hostage workflow,” not a whole company

Example wedges:

• Pull AR aging + reconcile customer IDs + push to CRM

• Extract student attendance + completion data from an LMS + sync to reporting

• Read orders from an old desktop app + update inventory planning

Keep it narrow. You want predictable UI paths, not infinite autonomy.

2) Build governance like it’s the product

If you don’t have:

• permissioning (who can the agent act as?)

• approvals (what needs sign-off?)

• audit logs (what happened, when, why, and what changed?)

…you don’t have an enterprise product. You have a demo.

Gartner explicitly ties agentic AI success to risk controls and governance—lack of those is a key reason projects die. Source

3) Treat “field reconciliation” as the core hard problem

Everyone wants to talk about models.

Your customer wants:

• “Is this the same customer?”

• “Which ID wins?”

• “What do we do when the UI value and export value disagree?”

• “How do we handle exceptions without a ticket storm?”

This is where you earn retention.

4) Make reliability boring

UI automation breaks. Popups happen. UI labels change. Sessions expire.

Your job is to:

• detect drift

• retry safely

• escalate cleanly

• never silently corrupt data

What system-of-record founders should do right now (before agents eat your lunch)

If you’re building (or running) a system of record company, this is your “adapt or die slowly” moment.

1) Stop thinking your moat is the database

Assume the customer can (eventually) extract anything.

Compete on:

• depth of workflow

• time-to-decision

• outcome ownership

• trust + compliance

2) Become agent-native instead of agent-hostile

If your system fights automation, agents will route around you.

Offer:

• stable APIs where possible

• export endpoints that don’t suck

• event streams / webhooks

• permission scopes that map to real roles

• built-in audit trails customers can reuse

3) Turn your product into a “safe execution environment”

If agents are going to act, the system of record can still win by being:

• the most trustworthy place to execute changes

• the best place to enforce policy

• the best place to validate actions before write

Basically: “You can do anything… but only safely through us.”

That’s a real moat.

“By 2028…” is way too slow (and also kind of the point)

Gartner’s numbers (33% of enterprise software includes agentic AI; 15% of daily decisions autonomous by 2028) are framed as forecasts. Source

But the more important detail is why Gartner thinks adoption will be slower than the hype cycle:

Because production-grade agents require:

• governance

• controls

• reliability

• real integration into messy workflows

That’s not a model problem. That’s an operator problem.

Good news: that’s where founders with battle scars win.

Bringing it home

Most companies don’t have a “data problem.”

They have a hostage situation.

And agents are the negotiation team that can finally get the data out—without waiting for the vendor’s roadmap, without a 9-month integration, and without Steve in accounting doing ritual exports every Monday.

Either be the AI company or get eaten by it…

Last year, we brought hundreds of the top Vertical AI & Vertical Software Founders to Miami for an unforgettable few days...

We heard biz and founding stories, playbooks, etc. from some of the most prolific Vertical Software Founders & CEO’s.

This years line up is INSANE. I actually can’t believe it. We have a bunch of unicorn founders coming to share hard earned lessons.

I can’t wait to share the speaker lineup with you all...

Sign Up To Be Notified When Tickets Drop

PS - I will give free tickets to a bunch of folks that opt in above :-))

Advertising Is Moving From Attention To Action. And You Can Cash In On It.

For the last 20 years, ads have mostly lived in places where people browse: feeds, search results, content pages.

But we’re walking into a world where people don’t browse as much — they delegate.

They ask an agent to “handle this,” the agent decides what to do next, and then the work gets executed inside the software that runs an industry.

That shift quietly turns “advertising” into something way more powerful: monetized distribution inside workflows.

And vertical SaaS is sitting on the best real estate on the internet.

Alright, let’s get to it.

OpenAI is testing ads inside ChatGPT
(and that’s the starting gun)

As of Feb 2026, OpenAI has begun testing ads in ChatGPT in the U.S. for logged-in adult users on the Free and Go tiers, while keeping Plus/Pro/Business/Enterprise/Edu ad-free. This is not rumor anymore — OpenAI published the test details and the product principles they’re using to keep trust intact.

OpenAI is being extremely direct about the rules of the road: ads are clearly labeled and visually separated, ads don’t influence the answers, and conversations are kept private from advertisers. They’re also shipping real user controls (hide/report, “why am I seeing this,” personalization toggles, delete ad data). That’s important because it tells you how serious “trust” will be as this spreads into work software.

Here’s what the format looks like right now — a sponsored unit underneath the response, with clear separation and labeling:

And yes, the early pricing chatter is aggressive. Reporting on the pilot said OpenAI confirmed a $200,000 minimum commitment, and buyers were quoted around $60 CPM. Whether those numbers hold long-term doesn’t matter as much as the fact the channel exists.

Once ads work in conversation, the next step isn’t “more ads.” It’s ads inside actions.

Chat ads are the obvious first move because they look like search ads: user expresses intent in text, platform shows something relevant.

But agents don’t stop at answering.

They recommend what to do next and often do it for you.

So the next logical ad surfaces aren’t “below the response.” They’re inside:

• the recommended next step

• the default tool choice

• the automated workflow the agent triggers

In other words, ads move from “attention” to “execution.”

That’s why “embedded ads” is a misleading term. What’s coming is closer to sponsored distribution inside the operating system of an industry.

Vertical SaaS has the unfair advantage: it knows intent, timing, and outcome

In horizontal consumer land, you can infer intent (“maybe they’ll buy shoes”) and you can try to attribute outcome (“we think the ad influenced a purchase”).

In vertical SaaS, the product already has the best signals on earth:

The software knows what job the user is trying to do, the exact moment they’re trying to do it, and whether it worked.

A restaurant platform knows when payroll is due. A construction platform knows when a change order is approved. A logistics platform knows when a shipment needs freight. A medical workflow knows when an order is being placed.

That’s a completely different level of “intent” than a keyword.

If you want a parallel mental model, think about what embedded finance did to vertical SaaS: instead of monetizing “traffic,” vSaaS monetized the workflow. This same shape will happen with sponsored actions.

The agent-as-dispatcher model (where this gets really interesting)

Imagine your vertical product has an AI layer on top — not a chatbot bolted onto the side, but an agent that can actually dispatch work.

A user says: “handle this.”

The agent replies: “Got it. Here are the best ways to do that.”

Then it offers two or three options. One might be “sponsored,” clearly labeled. The user picks. The workflow runs. You take a cut.

“Run payroll via X.”
“Finance this invoice via Y.”
“Order parts via Z.”

That’s not “banner ads.”

That’s the equivalent of being the default distribution layer inside a vertical economy.

And if you’re a founder, you should feel your stomach drop a little, because this is one of those shifts that creates $10B companies and also deletes whole categories of incumbents who used to own distribution.

The hard part isn’t building the model. It’s building trust without killing conversion.

If you’re going to put sponsored actions inside workflows, you’re now playing with production systems and money movement. That means your “ads product” needs to look more like enterprise infrastructure than a media platform.

OpenAI is already laying down the pattern: clear labeling, separation, privacy boundaries, and user controls. If you don’t match that bar (or exceed it in a regulated vertical), you’ll get one of two outcomes: either users revolt, or regulators do it for them.

And here’s the non-negotiable: if sponsorship changes what the agent “believes” is best for the user, you’re done. The only sustainable version is sponsorship that changes placement (and maybe incentives), while the agent remains outcome-optimized and transparent.

OpenEvidence is an early case study:
Vertical AI + Ads can scale fast

CNBC reported that OpenEvidence topped $100M in annualized revenue, and that it was one of the first AI startups to rely on advertising for revenue (rather than pure seat-based subscriptions). That ad-led monetization is a huge signal: in vertical contexts, “ads” can be positioned as “distribution to clinicians,” not as random brand spend.

Their own announcement highlights just how vertical this is: verified clinician usage, point-of-care workflows, and agentic research (DeepConsult). Whether you agree with every claim or not, the structure matters: a vertical AI product with heavy daily workflow intent can support monetization models that look “ad-like” while behaving more like outcomes-based distribution.

What founders should do next…

If you’re building vertical AI agents or vertical SaaS, the play isn’t to slap ads into your UI and hope. The play is to identify the handful of workflows where the user is already making a high-stakes tool choice — then let the agent become the dispatcher for that choice.

Start where partners already want distribution: payments, financing, procurement, insurance verification, staffing, parts ordering, logistics booking. The reason these are so attractive is because you can measure the full loop:
suggestion → selection → execution → outcome.

Then be boring about trust. Make sponsorship obvious. Give users control. Keep an audit trail. If you can’t explain why an action was recommended, you’re building a liability, not a platform.

And if you do it right, you don’t just add a revenue stream. You create a new moat: you become the place where decisions turn into executed workflows — and partners will pay endlessly to be in that decision set.

Have a product or service that would be great for our audience of vertical SaaS founders/operators/investors? Reply to this email or shoot us a note at ls@lukesophinos.com

The New Vertical Software Era → Storage vs. Action

Dave Pandullo and Scott Hoke of AQL Growth join Verticals to unpack why the next wave of vertical software isn’t about storing data — it’s about deciding what happens next and doing it.

Remember folks, the full recording is over on YouTube. Check it out there, and do us a favor if you can and subscribe while you’re there :-) Would mean a ton to us!

Watch Episode Here

Meet The Guests

Dave Pandullo and Scott Hoke have spent the better part of their careers doing one thing: partnering with founders who are building core software systems for niche industries. Both are general partners at Frontier Growth, a growth equity firm that’s been around since 1999 and has been investing in vertical software since 2007 — back when most investors thought vertical markets were too small to matter.

Their first vertical SaaS bet was a company called Vaxco, a club management platform for YMCAs and JCCs. At the time, they thought the market was tiny and got out a little too soon. That business went on to become a dominant platform in broader sports and recreation, growing north of $250–300 million in ARR. It was a lesson in the power of niche dominance — and one that shaped everything that came after.

Now, as part of a planned succession from Frontier’s founders, Dave, Scott, and another partner are launching a new firm: AQL Growth. The name comes from Aquila — Latin for eagle — a nod to the sharp vision and intense focus that defines their investment strategy. AQL is exclusively focused on growth-stage vertical software and vertical AI: companies that have established product-market fit, typically at $2–3M+ ARR, and are ready to accelerate into category leadership.

Scott, who’s been a growth equity investor for over 20 years, described the work with characteristic enthusiasm: “We get to sit on the ground floor and help companies with capital and expertise to see whole companies and all the employees just enjoy growing.” His background is in investment banking and finance, but it’s clear his passion lives at the intersection of building and investing.

For founders evaluating AQL as a partner, the key details are straightforward: they write $5M–$30M equity checks, primarily in primary capital but with flexibility for secondary components (cap table cleanup, founder liquidity). They’re minority investors only — they don’t acquire, they partner. Most of their portfolio companies are growing 50–100% at the time of investment, and over the past 12 months they’ve done eight investments spanning the trades, sports and entertainment, state and local government, and outpatient healthcare.

The New Vertical Software Era → Storage vs. Action

The phrase “system of record” has been the north star of vertical SaaS for nearly two decades. The pitch was simple: become the place where a business stores its most critical data, and you become impossible to rip out. Switching costs were your moat. The data was sticky because moving it was painful.

That logic is cracking. As Nic pointed out during the conversation, the transportability of data is now trivial. He shared a personal example: he recently switched an internal system of record by having AI harmonize and migrate all the data for him. Easy. If the switching cost was your primary moat, you’re in trouble.

“A system of record stores what happened.
A system of action decides what happens next — and then does it.”

— Scott Hoke, GP, AQL Growth

Dave and Scott’s thesis at AQL centers on what comes after the system of record: systems of action. Where a system of record captures data, a system of action governs the workflows where decisions are made and finalized — where work is dispatched, claims are submitted, revenue is recognized, and compliance is enforced. It’s where the actual jobs to be done for a business happen, increasingly with AI enabling those activities autonomously or with a human in the loop.

Luke drew a useful historical parallel. When he was running his company, they built an industry-specific BI tool with a notification engine behind it — sending texts or emails when metrics trended the wrong direction. That was pre-AI systems of action: it could only do one thing. What’s happening now is fundamentally different. AI can execute multi-step processes: create the invoice, send the invoice, run a five-step follow-up cadence to get payment, and reconcile back to the general ledger. That’s not a notification — that’s transformative automation.

The group landed on an important nuance: the system of record isn’t dead, but it needs a new source of defensibility. Nic framed it as a new layer of data that didn’t exist before — what some are calling the “context graph” or “decision layer.” It’s not just what happened, but why it happened. Why did a deal move from one stage to another? Why did you reach out to this contact instead of that one? That proprietary context, accumulated over thousands of decisions within a vertical workflow, becomes the new switching cost.

“CRM data — all my stuff is in there — that was why it was so sticky. But the new modes are closer to the workflow. The actions being taken and the data created as a result — that combination is sticky, and the basis by which you train models to take the next step.”

— Scott Hoke, GP, AQL Growth

Watch Episode Here

Today’s episode is brought to you by Parafin - the leading embedded capital partner.

• 4.8 star rating on Trustpilot, with over 500 reviews [source]

• 71% CSAT rating across all partners

• 84 NPS

• Serving top Fortune 500 companies, including Amazon, Walmart, and DoorDash.

Healthcare: A Living Laboratory

Much of AQL’s portfolio sits in outpatient healthcare — the world outside of hospitals and health systems where Epic dominates. They’ve invested in practice management platforms across allied health (chiropractic, acupuncture, mental health), veterinary, durable medical equipment, and home care. Across all of these settings, they’ve seen a proliferation of AI transcription and scribe tools. Their belief is that the practice management system is uniquely positioned to embed those capabilities rather than lose them to standalone point solutions.

But what excited Dave and Scott more than transcription was the opportunity in revenue cycle management — claims processing, billing, denials management. Historically, RCM was a services business that software investors viewed skeptically. You wanted pure SaaS, not services. But in the age of AI, Dave argued, the calculus has changed: “We’re big believers in delivering more value to the customer, automating things that are painful for them, and letting them focus on what they do best — which in healthcare is practicing.”

They pointed to Access Care, their investment in the non-medical home care space, as a case study in domain complexity creating defensibility. Access Care coordinates caregivers who deliver services to elderly or disabled people in the home. Those services are reimbursed by Medicaid, which operates differently in every state. Every visit has to be governed by care plans, authorization caps, licensing requirements, overtime rules, and electronic visit verification to prevent fraud. Reimbursement rules are constantly changing. This isn’t a simple scheduling problem — it’s a deeply complex enforcement equation where new data is constantly flowing from caregivers, coordinators, and third-party sources. That complexity is the moat.

Is the System of Record Actually Safe?

Luke played devil’s advocate with a question that’s been circulating in the ecosystem: there’s a lot of posts going around saying “SaaS is dead, but if you’re the system of record, you’re safe.” Based on the thread the group had just pulled — that agents can extract data, create digital twins, and trivialize migration — is the system of record actually safe?

Dave was direct: “I don’t necessarily agree that under every circumstance, the system of record is safe. Vertical-specific CRMs are definitely at risk.” But he drew a sharp distinction between shallow systems of record (essentially searchable databases) and deep, domain-complex platforms where the system of record is embedded in regulatory workflows, multi-party data flows, and industry-specific compliance.

The Slice example resonated deeply with the group. Luke brought up Aleer at Slice — who they featured in their first Verticals episode — as someone who got this right before everyone else. Aleer took what everyone assumed was already commoditized (the POS for pizza shops) and wrapped moat after moat around it: services, ordering, supplies, and now actual delivery infrastructure with vans. He built a “digital franchise” — understanding every aspect of his customer’s business better than the customers themselves.

“Domain knowledge is built into the workflow, into the platform, and that is really hard to replicate.”

— Scott Hoke, GP, AQL Growth

The group debated why more founders aren’t pursuing the “franchise in a box” model. Scott attributed it to maturity — the AI revolution is still too young, and most operators running fragmented SMB verticals are focused on running their businesses, not building software with agentic AI at the forefront. Dave added that prioritization plays a role: even in this new wave of entrants, founders are still focused on delighting customers with their core value prop. He pointed to a portfolio company in orthotics and prosthetics that’s just now launching automated purchasing integration with manufacturers — the bureaucracy and logistics required to deliver that end-to-end are staggering.

Nic offered perhaps the most practical insight: Slice didn’t start as a franchise-in-a-box. It started at $2 per order — a low-friction wedge that built trust. The “10% of your revenue” conversation is intimidating; the “$2 per order” conversation is not. That wedge strategy, especially in SMB industries where customers don’t know or care about AI, might be the most powerful starting point for founders who want to eventually wrap their software around entire businesses.

Eagle Vision: How AQL Built Their Own System of Action

One of the most compelling parts of the conversation was Scott’s candid walkthrough of Eagle Vision — the internal AI platform AQL has built to transform how they source and evaluate investments. It’s a powerful case study of practicing what you preach.

It started, as Scott tells it, with a humbling realization. A couple of years ago, he began experimenting with AI for deal sourcing: “I tried to say, is this company a good target for our investment thesis? And the response was just... it completely whiffed.” The outputs were garbage because they weren’t being specific enough about what they actually believed. That failure sparked something important — it forced AQL to sharpen their thesis to a degree they never had before.

The journey began with a CRM audit. AQL started thinking of themselves as a product company — capital is increasingly commoditized; it’s a product. So who is their ICP? How many targets exist? How should they be ranked? When they dug into their CRM (HubSpot), they found what Scott diplomatically called “a filing cabinet with the wrong information in it.” Companies were mislabeled. Great targets had been passed on because someone in a hallway said “we don’t like marketing automation” or “construction tech is too cyclical” — tribal knowledge misapplied at scale.

“I called Dave and said: I just scanned 25,000 software companies with AI according to our thesis. Our TAM is 6,500 vertical software companies fitting our strategy.”

— Scott Hoke, GP, AQL Growth

Dave emphasized a crucial point: Eagle Vision wouldn’t have been possible without a laser-focused strategy. If AQL invested across dozens of sectors and stages, the permutations of prompts would have been infinitely more complex. Their narrow focus on vertical software made it feasible to encode their judgment comprehensively. “It flipped the whole model on its head,” Dave said. “Old-school growth equity had analysts screening manually, cold-calling at the top of funnel. Now we’re getting partners on planes before we’ve even had a conversation.”

Scott was refreshingly honest about the process: “This didn’t speed us up initially. It slowed us down. It forced us to get sharper on our thesis, to look at the outputs and analyze how accurate they were.” They used multiple AI models for different purposes — some for web research, others for content creation — and finding the right combination was trial and error. The tech stack includes coding tools like Cursor, platforms like Lovable and Supabase, Clay for data enrichment, and HubSpot as the CRM backbone.

Case Study — Albie

Eagle Vision identified Albie, a business management platform for damage restoration providers (a subset of the trades where ServiceTitan’s ecosystem stops). The company scored highly across AQL’s thesis dimensions despite non-standard retention metrics that other investors flagged. When AQL dug deeper, they found the retention issue was driven by ICP cohort segmentation — the core customer base had incredible retention and customer love. AQL moved fast, shared deep vertical knowledge with founder Alex Duda, and won the deal. Retention has “dramatically changed” in the nine months since investment.

Albie itself is becoming a system of action in restoration. At a recent board meeting, the entire discussion centered on how they’re delivering discrete AI use cases — specific bots, each solving a concrete pain point — rather than talking about AI at a high level. As Dave explained: “Individuals that run businesses in that industry don’t want you to talk high level. They want you to solve a pain point for them.” Albie is shipping three tested AI bots per quarter, with a vision of progressively automating back-office and even field operations so restoration businesses can focus on generating leads and getting paid faster.

Watch Episode Here

Eating Your Own Cooking

One theme that ran through the entire conversation was the importance of walking the walk. Dave referenced Kyle Norton’s talk at the Vertical Software Summit: if you’re not AI-native and you’re not hiring AI-native talent, there’s a big problem. AQL took that to heart, hosting a portfolio CEO summit entirely focused on AI initiatives, which sparked cross-portfolio collaboration and idea pollination.

“AI didn’t make us lazier. People that are truly hungry and curious — it makes you busier. The art of the possible has my mind racing.”

— Scott Hoke, GP, AQL Growth

Scott’s closing advice was personal and energizing: “Take a day or two — a sabbatical — and dive into these tools. Ask questions. You’ll go on a journey. If you’re inquisitive, hungry, and humble, this is the era for you.” It’s advice that clearly comes from someone who has lived it. Dave called Scott “definitely AI-native” — and Scott’s response revealed the quiet joy of someone who’s genuinely excited by what they’re learning: “I’m like going back to school for the first time.”

The conversation closed with a rapid-fire section that revealed some telling details. When asked about his favorite AI wedge product, Dave pointed to Clay — not vertical, but incredibly powerful for ICP data enrichment, and something they’re actively showing to portfolio companies. When asked about the biggest remaining data gap, Scott landed on private company revenue data. The old proxy of employee headcount-to-revenue ratios is “completely bunk now” in the AI era: a company with four employees could theoretically have hundreds of millions in revenue. If someone could create a reliable source of private company revenue data, Scott argued, it would be enormously valuable.

Key Takeaways for Founders

Distilled from the conversation — actionable insights for vertical software builders navigating the AI transition.

Watch Episode Here

Have a product or service that would be great for our audience of vertical SaaS founders/operators/investors? Reply to this email or shoot us a note at ls@lukesophinos.com

Today’s newsletter is sponsored by Xplor Pay. In a case study they show how a POS provider embedded payments using PayFac as a Service without taking on the burden of compliance, risk, or merchant onboarding.

Learn how white-labeled payments, streamlined setup, and flat-rate pricing helped the platform improve the experience while unlocking new recurring revenue. Read the full case study here.

Alright, let’s get to it…

SaaS Company? Think About Adding An AI Agent App Store This Year

Every SaaS company that is 4 years old or more is racing to add AI products and transform its business into an AI Company.

At this point, I think most realize that
A) they need to do this from a valuation perspective and
B) the world is headed in this direction.

I do want to propose something to think about for platform companies that find themselves in this exact circumstance…

Should you add your own AI Agent products? Yeah, I think so. I sure would be. Hopefully you already have, but if not, I’d get on it.

But MORE IMPORTANTLY, there are other creative ways to drive value and position yourself well in this new era.

What about building the AI Agent App Store for your industry?

What if instead of just relying on yourself and your team to add net new products, you go partner with every upstart in the industry and launch an AI Agent App Store.

This is nice for a few reasons…

#1. You're not solely dependent on your internal resources to build/create/ship/iterate/etc
#2. You add a nice platform moat. If you are a system of record for your industry, this strategy will make it so everyone can play with / test AI products directly within your tool, leveraging your system’s data set.
#3. You won’t be on the HOOK for every AI product. Similar to the iOS App Store, folks aren’t blaming Apple for bad Apps. They're blaming the App maker themselves…

I haven’t seen ANY industry Systems of Record of Platforms take this approach, and I think it’s a really smart one to create staying power. You can still build your own AI tools, just launch them in YOUR app store.

App Stores are beautiful moats and beautiful business models. Nothing beats a toll fee.

A few REAL WORLD Examples of companies that SHOULD do this…

1. ServiceTitan (for home services/trades like plumbing, HVAC, electrical)
   
   ServiceTitan is a comprehensive system of record for trade businesses, centralizing job scheduling, customer records, invoicing, dispatching, inventory, technician performance, and financials. It holds deep proprietary data on job histories, customer interactions, pricing, and operational metrics.
   
   If they did this: They could launch a “ServiceTitan AI Agent Marketplace” where third-party AI developers build specialized agents (e.g., predictive maintenance agents that analyze historical job data to forecast equipment failures, dynamic pricing agents for quotes based on real-time market and customer data, or automated follow-up agents for customer satisfaction). These agents would run natively in ServiceTitan’s dashboard, accessing the platform’s dataset securely via APIs.

   ServiceTitan takes a cut (e.g., 20-30% toll) on transactions or subscriptions, while avoiding blame for underperforming agents and accelerating AI innovation without building everything in-house.

2. Toast (for restaurants)

   
   Toast is the dominant point-of-sale and management platform (system of record) for restaurants, managing orders, menus, inventory, staff scheduling, payments, customer loyalty data, sales analytics, and online ordering integrations. It captures granular, real-time data on every transaction, menu performance, and guest behavior.

   
   If they did this: A “Toast AI Agent Store” could partner with AI upstarts for agents like intelligent menu optimizers (analyzing sales data to suggest dynamic pricing or item removals), predictive staffing agents (forecasting busy periods from historical trends and external factors like weather/events), or personalized guest experience agents (generating tailored recommendations or automated re-engagement campaigns). Agents would integrate directly into Toast’s POS terminals or back-office app, using the platform’s rich transactional dataset. Toast monetizes via app fees or rev-share, builds a stronger moat by making Toast the central hub for restaurant AI experimentation, and lets developers iterate fast on niche tools.

3. Mindbody (for fitness/wellness studios, gyms, spas, and boutiques):

   
   Mindbody is a leading system of record for the wellness industry, handling class/instructor scheduling, client bookings, memberships, payments, attendance tracking, retail inventory, and marketing for studios and spas. It stores extensive client profiles, booking histories, class performance, and revenue data.

   
   If they did this: They could create a “Mindbody AI Agent App Store” for partners to offer agents such as personalized workout planners (using client history and goals to suggest classes/programs), retention prediction agents (flagging at-risk members from attendance patterns), or dynamic pricing agents for classes/memberships based on demand data. These would embed in the

   Mindbody app or client-facing portal, leveraging the platform’s dataset for hyper-relevant AI. Mindbody earns platform fees, reduces reliance on internal AI builds, and positions itself as the indispensable ecosystem hub—much like the blog’s vision—while third-party developers handle specialized AI innovation.

Obviously this won’t be applicable to upstarts, but if you have a System of Record, EVEN A SMALL ONE, this is a competitive advantage…

So at least think about it…

Last year, we brought hundreds of the top Vertical AI & Vertical Software Founders to Miami for an unforgettable few days...

My favorite part?

I’ve now heard THREE separate stories of Founders that met investors here and raised subsequent rounds with them shortly after.

How cool is that???

Magic happens when you get hundreds of incredible Vertical AI & SaaS founders in a room for a few days...

We’re doing it again, but Bigger & Better in H2 2026 😃 😃 😃

I can’t wait to share the speaker lineup with you all...

It’s truly a who’s who of VERTICALS.

Sign Up To Be Notified When Tickets Drop

PS - I will give free tickets to a bunch of folks that opt in above :-))

Why Casca Could Be a Key Player in Vertical AI for Lending

Most Vertical AI companies don’t win because the tech is clever. They win because the economics are impossible to ignore.

Casca (formerly known as Cascading AI) is an interesting example in the Vertical AI space right now—a case study for why some wins aren’t just about flashy AI tech—they’re about potentially rewriting the economics of an industry.

I’ve been following lending tech for a while, and the pain in loan origination is brutal. Legacy systems—think COBOL-era software from the ‘80s or ‘90s—are slow, manual, error-prone, and completely incapable of scaling in today’s environment. For banks (especially community and regional ones) and non-bank lenders, this translates to real money bleeding out:

• Sky-high cost per loan processed

• Weeks-long decision times that kill applications (abandonment rates are painful)

• Lost revenue from deals that walk away to faster fintechs or worse—predatory lenders charging insane APRs

• Compliance headaches baked into outdated, fragmented workflows

Casca’s insight cuts through the noise: don’t just sprinkle AI on top of broken processes. Build an AI-native loan origination system from the ground up that replaces the legacy core. If you modernise the core lending workflow, you don’t just improve efficiency, you unlock growth.

This isn’t augmentation—it’s full rip-and-replace for mission-critical infrastructure. Banks aren’t experimenting; they’re embedding Casca into their lending stack because it directly ties to revenue throughput, lower costs, and better compliance at scale.

I’ve been digging deeper into Casca since spotting their early momentum, and the more I peel back the layers, the clearer it becomes why this could be one of the more compelling Vertical AI plays in fintech right now. It’s not just about the AI hype—it’s about a team that’s laser-focused on ripping out legacy pain in a market where the economics scream for disruption.

The Founders: Banking Insiders Meet AI Expertise

Casca was founded in 2023 by Lukas Haffer and Isaiah Williams, a duo that brings the perfect blend of domain depth and technical chops to tackle loan origination’s messiest problems.

Together, they’ve assembled a 37-person team (as of late 2025) heavy on engineering and product talent, with advisors including FDIC bank CEOs and fintech scouts like Simon Taylor (a16z scout and Fintech Brain Food founder). This isn’t a generalist AI startup—it’s vertical from day one, born from insiders who know banking’s underbelly.

“Legacy loan systems aren’t just inefficient; they’re revenue killers. By going AI-native, Casca isn’t augmenting—it’s rebuilding the core to unlock trillions in trapped value.”

Traction: Real Metrics in a High-Stakes Market

Casca’s early traction is impressive, especially in a space where enterprise adoption moves at glacial speeds. They’re already embedded with flagship customers:

These aren’t pilots; they’re production integrations driving measurable outcomes:

• Commercial loans processed 10x faster than other fintechs and 30x faster than industry averages, with prequalification in just 5 minutes

• Automates 90% of manual effort, handling 10,000+ pages of financial docs, 40+ KYB checks, and instant ratio calculations via 30+ native integrations

• 3x higher conversion rates through digital apps and auto-prefill; a 24/7 AI Loan Assistant responds in 2-3 minutes

• Platform trusted by FDIC-insured banks and non-banks alike, setting new standards for SBA and commercial lending

Funding: Betting Big on Infrastructure Overhaul

The funding story tells the same tale. Casca raised $3.9M in pre-seed back in early 2024 to prove product-market fit. Then, in August 2025, they closed a massive $29M Series A—led by Canapi Ventures, with participation from Live Oak Ventures, Huntington, Peterson Ventures, Y Combinator, and others—bringing total funding to $33M.

That’s not “point solution” or “nice-to-have” money. That’s capital betting on core system replacement in a multi-trillion-dollar global lending market.

“Casca stands out... They’ve worked alongside top AI researchers and within banks themselves to simplify business lending using responsible AI.”

— Neil Underwood, Canapi Ventures

The business model aligns perfectly with enterprise reality in financial services:

• High ACV (tied to loan volume and throughput)

• Long sales cycles, but sticky, long-term contracts once embedded

• Extremely high switching costs (you’re replacing the heart of origination)

• Buyers with real budget authority (CFOs, heads of lending at banks/credit unions)

Why This Matters for Vertical AI

Casca sits in that rare spot: mission-critical workflow + direct revenue linkage + capital aligned with infrastructure overhaul, not experimentation. Their story reinforces an interesting take:

Will the biggest wins in Vertical AI come from full system replacement or bolt-on augmentation?

With mission-critical ties to revenue (more loans, lower costs, better compliance), high ACVs linked to volume, and sticky contracts in a regulated space, the moat is massive. They’re not selling demos—they’re selling economic transformation in a market ripe for it.

It’s a strong reminder in Vertical AI: the real winners aren’t the ones with the cleverest prompt engineering or the shiniest agent demo. They’re the ones making the economics impossible to ignore—by going after structurally broken, high-cost processes and replacing them entirely.

So, the question I keep asking myself:

Do the biggest Vertical AI wins come from augmentation or full system replacement?

Casca’s take is its full system replacement that unlocks the outsized returns. Augmentation is valuable (and easier to sell early), but it rarely changes the unit economics enough to dominate. When you rip out the legacy core and build AI-native from the start, you capture the full value—efficiency, speed, scale, and defensibility.

Casca is proving that thesis in real time in one of the toughest verticals out there. Excited to see how far they take it. Watching this one closely as they expand.

If you’re in fintech or Vertical AI, what’s your read—can rip-and-replace scale faster than augmentation in regulated industries?

Have a product or service that would be great for our audience of vertical SaaS founders/operators/investors? Reply to this email or shoot us a note at ls@lukesophinos.com