Are AI-native service companies VC-backable?

Two definitions before we start, because most of the noise on this topic comes from people using the words loosely.

An AI-native service company delivers a traditional service... still through human professionals... but rebuilt so that AI does most of the work and the humans do the part that actually needs judgement. VC-backable means something narrower and harder than "a good business". It means a company that can grow fast, hold strong margins, and exit at a premium multiple. Venture-shaped, not just profitable.

Keep both of those in your head, because almost every argument about this category falls apart the moment someone blurs them.

Why this is suddenly a question

Five years ago this article could not have been written, and not because VCs were short-sighted. The models simply could not do the work. You could automate the edges of a service... scheduling, formatting, a bit of search... but the core, the part a client actually pays for, still needed a human for every unit of output. That capped you at a traditional services business: linear, people-bound, 30% margins on a good day.

What changed is raw capability. The current generation of models can carry a meaningful share of skilled professional work end to end, not just assist with it. That is the whole game, because it is the first time you can grow revenue in a services business without growing headcount at the same rate. Decouple those two lines and a services company starts to behave like software: high margins, real operating leverage, scale that is no longer bounded by how many people you can hire and train. That single shift is what turned "services are un-backable" into "services might be the biggest opportunity in venture". Everything else here follows from it.

The tide turned, but not where you think

Twelve months ago, pitching an AI-native services company to most VCs was close to impossible. "Services" was a slur. Linear scaling, people-heavy, low multiples, no thank you. I lived that wall personally.

Then the thesis got written down. Sequoia's Julien Bek published "Services: The New Software". Emergence's Gordon Ritter, who had been circling the same idea since 2016, put it into "Above the Model". YC devoted a Startup School session to it. a16z and Bessemer started committing real money to it.

And more important than the essays... the bets started landing.

But that money is crowding into a handful of proven names, and that is where it gets misread. The public conversation and the private one have drifted apart. On LinkedIn it is settled: AI-native services are the next big thing, obviously, everyone always knew. In the conversations I actually have with investors, backing a new, unproven AI-native services company is still a genuine debate. The gap between the public hype and the private caution is wide, and most of the interesting truth lives in that gap.

This article is about that gap.

Why VCs actually care: the markets are vast

Here is the part that makes a VC lean forward. For every dollar spent on software, several are spent on the services around it. The tool markets are big. The work markets are enormous. That is the whole appeal, and nowhere is it clearer than in legal right now, where the smartest funds are effectively betting against themselves.

On one side, the copilots. Harvey raised $200m at an $11bn valuation, on roughly $190m of ARR... a multiple near 58x... with more than 100,000 lawyers on the platform and over $1bn raised in total. Legora raised $550m at $5.5bn, tripling its valuation in months. Both sell the tool to law firms.

On the other side, the autopilots. Crosby and Lawhive are not law firms with AI bolted on... they are tech companies that employ lawyers and sell the finished work. Norm AI went a step further and launched Norm Law, an AI-native firm for financial institutions, bringing in the former chairman of Sidley Austin as its chairman. Same Tier 1 names, opposite bets. Sequoia backs Harvey and backs Crosby. Bain and Coatue are on both sides too.

Why would the most sophisticated money fund both sides of the same fight? Because the legal-tech market is small... around $36bn... while the legal services market is around $1.1 trillion. An $11bn valuation on $190m of revenue only makes sense if you are not really buying a software company. You are buying a call option on who becomes the biggest law firm of the future. Get inside every firm first, own the workflow, then own the client relationship. It is the same move the model labs are running... start as the infrastructure everyone builds on, end as the product that competes with them. Harvey starts as the tool. It may well end as the firm.

That is the real reason services went from red flag to thesis. Not because the work is glamorous, but because the markets underneath it are many times larger than software, and AI is finally the lever that takes software margins out of them.

First, the success stories... because they are real

Let me give the optimists their due, because the wins are genuine and most of these companies are barely out of the gate.

Legal. Crosby, the hybrid AI law firm that returns a marked-up contract in under an hour, now sitting on a $60m Series B. Lawhive, which became the firm rather than selling software to one, compressing legal work for consumers and SMEs. And Eudia, taking the enterprise route with $105m from General Catalyst and its own in-house legal delivery team.

Insurance. Corgi is the one I would watch most closely... a licensed, full-stack, AI-native carrier that has now raised over a quarter of a billion dollars and crossed a unicorn valuation. Harper, the AI-native commercial brokerage that is the broker rather than a tool for brokers. WithCoverage alongside it. Strala and Pace on claims.

Everywhere else. Arca just came out of stealth in wealth management with $64m and General Catalyst leading. Panacea in FDA regulatory work. Jack & Jill in recruitment. M&A Research Institute in Japan, which made its founder a billionaire by compressing deal timelines on exactly the succession wave we are riding.

I could stop right here, point at that list, and write "yes, AI-native services are VC-backable". It would be true and completely useless. Because the honest answer is that it depends entirely on the service you are providing.

So let us build the test.

The Masquerade Test

Sequoia's line is that the next trillion-dollar company will be "a software company masquerading as a services firm". I've turned that idea into a test... the cleanest way I know to separate the backable from the un-backable. I call it the Masquerade Test, after Sequoia's line.

One question sits underneath all of it: are you genuinely the software company in disguise, or just a services firm wearing an AI costume?

Here is what the test actually asks:

1. Does your revenue decouple from your headcount?

YC put this better than I can. If your revenue grows in lockstep with how many people you hire, you have a problem. Not always... if one person is bringing in serious money, fine... but as a rule, your technology should be making each person more and more productive over time. Data flywheels, internal process automation, economies of scale across both employees and clients. Traditional advisory runs one partner, five associates, four deals. Eilla runs one senior advisor pair plus the AI layer across many more mandates than a boutique could ever touch. The heads do not move with the revenue. That decoupling is the first thing a VC is hunting for, and most "AI-native" companies fail it quietly.

2. Is the market big enough, and is there room for you?

The gold mine is opening a market that did not exist, or one that existed but was never economically reachable. For example, we are going after the hundreds of thousands of owner-led companies that need an advisor but are too small for a traditional firm to serve economically... a slice worth north of $50bn and almost completely untapped, precisely because the old economics never reached down there. If instead you are fighting for a market the incumbents already own, being a bit more efficient will not cut it, and VCs know it.

3. Are you actually AI-native, or a services firm with a bit of tech bolted on?

Be honest with yourself here, because the badge is cheap now and everyone is wearing it. Are you making the professional 10x more effective with real automation, or are you just one more shop with a Claude subscription? If you are an "AI-native" marketing agency using AI the same way every other agency uses it... I have bad news. Plenty of companies wear the AI-native badge while running an old-fashioned, people-heavy operation underneath. Maybe that is unfair in any given case, but it is exactly the right question, and it is the one you will get asked in the room.

4. How good are your margins, really?

Human capital is expensive. If you are fixing inefficiencies by throwing people at them, you are not AI-native, you are a staffing firm with a nicer logo. The opposite trap is just as real: if your token spend is so high that it would have been cheaper to hire three people, that is not a business either. Put model spend, hosting and every human-in-the-loop into cost of goods sold, not operating expenses where they quietly flatter the picture. If gross margin is not expanding as you grow, the AI is not pulling its weight, and a good VC will find that out before you do.

5. What is your moat against the incumbents, long-term?

Traditional firms are not stupid. The fact that law firms are run by lawyers and not engineers does not mean they will never wake up to how much money they are leaving on the table. They have brands and deep pockets, and they can copy your tech. So "our tech" is not a moat. The real moat is the compounding stuff: proprietary workflow, data flywheel effects, network effects, the process knowledge you only get by running the work yourself. One operator detail people miss... make sure your engagement terms even let you learn from the work. If they do not, you are building someone else's flywheel and you will not notice until it is too late.

6. How do you get to $1bn in revenue... and keep it?

This question rolls all the others into one. VCs need a path to enormous outcomes. Being profitable is lovely for you, but VC maths does not run on profitability, it runs on exits. And their deepest fear about you is that you are a traditional business in disguise that will never deliver software-like returns. If your growth is bottlenecked by hiring lots of people, you have already answered the question, and the answer is no.

7. What is your multiple at exit?

A big exit is big revenue times a big multiple. Get to $100m of revenue on a 0.5x multiple and you have built a $50m exit. No VC is excited by that. Premium multiples come from being more software company than services company... the tech, the scale, the data, the leverage. Treat the technology as something secondary and you have decided your own multiple in advance. This is the WeWork lesson in reverse: you can call yourself a tech company all you like, but the market prices you on what you actually are.

8. And then the part that has nothing to do with you: VC psychology.

Investors are human, with their own fears and scar tissue. For years, "services company" was an almost un-fundable phrase. So yes, there is hype now... but hype is not conviction, and conviction is what writes cheques. They want proof points. Give it a year. Once the first few hexacorns in this space post real revenue growth, the same funds that passed will tell you they "always believed in it". The truth is that venture is far more fearful than the name suggests, which is exactly why the committee will grill any new AI-native services company unusually hard right now. My honest advice: if you can, do not raise into this precise climate. Wait for the proof points to land. You will be in a stronger position and you will be able to create real competitive tension.

Should you raise at all?

Which leads to the question you should be asking before "am I VC-backable"... should I raise at all?

Great companies do not chase capital, they attract it. And the genuinely good thing about a real AI-native services business is that it is profitable by design, unlike the Ubers of this world. When you start winning, the VCs come to you. So the bar for raising should be high, not low.

There are plenty of stories of bootstrapped founders who out-earned VC-backed founders who over-diluted themselves into a corner. And it is not only about the money. There are just as many stories of founders who lost control and got pushed out by their own investors. Raising venture is not a free upgrade, it is a whirlwind with real downside. If you can win without it, my default answer is: win without it.

However.

Would your market even allow that? Picture a bootstrapped legal-tech founder raising slowly while Tier 1 funds pour money into Legora and Harvey. You are not raising slowly anymore, you are losing. I think the same dynamic is now playing out in AI legal services. You will not out-build a well-funded Crosby or Lawhive from a standing start. You simply will not.

You only win bootstrapped in two situations. Either you already dominate your market and have built proper moats, in which case why are you reading this. Or your market is so small that the big players and the VCs cannot be bothered to enter.. and if that is true, you were not VC-backable in the first place.

And there is a third path where the answer is almost always raise, however disciplined you are about dilution. In my last piece I argued there are three ways to build one of these companies:

• start as the software vendor and become your own client
• start with the domain expertise and add the engineering
• buy traditional firms and rebuild them with AI

That last one... the roll-up... is not a bootstrapping question at all. Acquisitions cost real money, and you cannot buy and transform a fragmented industry out of cash flow. If that is your path, you need outside capital by definition, and the good news is the capital exists and is hunting for you. There are firms purpose-built for it, like Tenet in Europe, and generalists going very big on it: General Catalyst has put $1.5bn behind its "Future of Services" roll-up strategy, writing nine-figure cheques to founders who acquire service businesses and re-rate their margins with AI. If you are rolling up, the question is not whether to raise. It is who to raise from.

So the two questions fold into one. Whether you are VC-backable and whether you should raise are the same question asked from two directions. The Masquerade Test answers both.

The honest summary

The thesis is right. AI-native services will be one of the defining business models of this decade. But "the thesis is right" and "your company is backable" are not the same sentence, and the difference between them is where most of the money will be made and lost.

The thesis is the easy part. Whether you pass the test comes down to the service you picked and how honestly you have built it. Are you the software company in the costume, or just the services firm wearing it? Be honest about the answer, because the market will be.

The proof points are coming, and when they land this stops being a debate. By then the winners will already be pulling away. The time to build the real thing is while people still doubt it.