The Doctrine Says: AINS Is Not SaaS. It's the Operator's Next Play

The TL;DR

Last week in New York, Eric Newcomer sat in on a private summit at Crosby, an AI-native law firm, hosted by Emergence Capital's Jake Saper. The topic: AINS, or "AI-Native Services." Not software that helps lawyers work faster. A law firm, an accounting shop, an insurance brokerage, built from the ground up on AI, charging for the completed contract instead of the billable hour. I read the piece twice. Then I read Saper's AI-Native Services Playbook. Here is my read: this is not a VC fad. This is the owner-operator's next play, and you do not need a term sheet to run it.

The Contrarian Angle: This Was Never a Venture Story

Everyone covering AINS is writing it as a funding story. Sequoia partner Julien Bek published "Services: The New Software" and it hit three million views. Bessemer built a three-part scoring framework for grading which services markets are "ripe." Emergence is out with a Playbook, a Summit series, and a cease-and-desist from Deloitte's lawyers after Saper's original piece, "The Death of Deloitte," argued the Big 4's $200B-a-year business was suddenly automatable. Every one of these firms is asking the same question: which AINS startup should we fund next?

Wrong question, if you already run a business.

You do not need Emergence's check. You need Emergence's insight, applied to the shop you already own. The insight is simple: services businesses used to be uninvestable because revenue scaled with headcount. AI breaks that link. Once it breaks, the business you already run, the agency, the bookkeeping practice, the claims shop, the staffing firm, stops being a "lifestyle business" and starts being an asset. That reclassification is the whole story. Founders chasing a Series A are optimizing for a pitch deck. You should be optimizing for a balance sheet.

I spent eleven years in the Navy before Hartford and Munich Re taught me how underwriters price risk for a living. Both jobs ran on the same discipline: you do not scale headcount to solve a capacity problem, you fix the system that created the capacity problem. AINS is that discipline, applied to professional services, forty years late. The doctrine is not new. The tooling finally is.

The Data: What "Software Margins in a Services Business" Actually Looks Like

Skip the pitch decks. Look at TripleDart. Bootstrapped in Bengaluru in 2020 with roughly $6,000 in seed capital (Rs 5 lakh), no venture money, ever. TripleDart now runs $7 million in ARR at a 50% EBIT margin, serving 300-plus B2B clients including GE, SentinelOne, and ByteDance, and managing more than $200 million in client ad spend with roughly 120 people. Founder Shiyam Sunder put it bluntly: "We rebuilt the function as software from day one. When a bootstrapped team can do that profitably, the 'agency' label stops fitting. That's a category, not an agency."

Compare that to the venture math. Bessemer notes professional services is roughly 13% of U.S. GDP, ten times the size of the entire software industry. Emergence pegs the addressable shift at $3 trillion. Rillet, an AI-native accounting infrastructure company, just raised a $100M Series C at a $1B valuation, its third raise in twelve months, on the strength of finance teams running "a tenth the traditional size." Legacy professional services firms run at roughly 20% gross margin. AINS operators are posting 50%+ on the same underlying labor.

Here is the number that should stop you: TripleDart hit that margin profile with zero dollars of outside capital. No board seat. No liquidation preference. No pressure to grow 3x a year to justify a $1B mark. Just an in-house AI platform, Slate, that automates SEO, content, and ad operations, plus the discipline to say what used to take five people now takes two.

| Model | Fee Basis | Typical Margin | Who Owns the Multiple | |---|---|---|---| | Traditional agency/services firm | Hours or % of spend | ~15-20% gross | Nobody, it dies with the founder | | VC-backed AINS startup | Outcome/completed unit | 50%+ gross | Investors, then founder, on an exit timeline set by the fund | | Owner-operator AINS rebuild | Flat retainer or per-outcome | 40-50%+ EBIT | You, on your own timeline |

The Case Study: TripleDart Is the Proof, Not the Exception

Saper's Playbook, built from interviews with Crosby's Ryan Daniels, Harper's Dakotah Rice, and Hanover Park's Chris Hladczuk, keeps returning to one operating principle: sit the people doing the work next to the people building the AI. Crosby's lawyers used to hand engineers batched feedback once a quarter. They switched to feedback every few hours, and the system improved fast enough that Daniels now talks about an "80-client" threshold, the point where pattern recognition replaces guesswork about which client requests are rules and which are exceptions.

TripleDart ran the same play without a Silicon Valley cap table. It stopped selling hours and started selling a flat monthly retainer decoupled from ad spend, "whether you spend $20k or $200k a month, our fee doesn't change." That single pricing decision forced the same discipline Emergence's portfolio companies discovered with venture money behind them: if your fee is fixed and your delivery cost is variable, the only lever left is automating the delivery. TripleDart's Slate platform is the compartment where that automation lives. Everything outside it, strategy, client relationships, judgment calls, stays human.

This is the tell that separates AINS from a rebrand. A services firm that bolts a chatbot onto its existing process is not AINS. TripleDart rebuilt inbound marketing as software first, then hired the humans to sit on top of it. Crosby did the same with contract law. The order of operations is the whole difference between an agency with an AI feature and an asset with a defensible margin.

The Sovereignty Stack: Own the Delivery, Don't Rent It

This is where the framework does its work. The Sovereignty Stack says the same thing in every domain: own your delivery infrastructure, do not rent it from a platform. Rent the delivery layer, and someone else sets your margin ceiling, your pricing power, and eventually your exit value. Own it, and you compound.

Applied to AINS, the Stack has three layers.

Layer one: the system. This is your Slate, your proprietary playbook, your codified way of doing the work that used to live in one senior person's head. If your delivery process only exists as tribal knowledge passed watchstander to watchstander, you have no asset. You have a job. Write the system down, automate the repeatable 80%, and you have created the thing a buyer can actually acquire.

Layer two: the pricing. Move from hourly to outcome. Per completed contract, per closed claim, per qualified pipeline dollar, whatever your version of "the fish, not the pole" looks like. Hourly billing punishes you for getting faster. Outcome pricing pays you for it.

Layer three: the exit. A services firm billing hours sells at 1-3x revenue, if it sells at all, because the buyer is really buying a Rolodex and a set of employment contracts that can walk. An AINS business with a codified system, outcome-based contracts, and margins that look like software sells at a software multiple, 5-10x revenue or more, because the buyer is acquiring a compounding engine, not a headcount.

Most owner-operators I talk to, in the $500K to $5M revenue range, are sitting on layer one already. They just never wrote it down or automated it. That is the whole first move: compartmentalize what you do into what a machine can run and what a human must judge, then build the machine.

The Honest Caveat: This Is Not a Guaranteed Multiple

I will not sell you the clean version. AINS carries real risk, and Saper's own Emergence Capital summit surfaced it directly. Founders at the NYC event debated whether to even mention "cheaper" in their pitch, because a fast, cheap AI-native service can read to a buyer as unreliable. The room mostly landed on: sell better, not cheaper. That is a warning for owner-operators too. If you rebuild your delivery as AI-native and lead with price, you will attract the wrong client and validate every skeptic who thinks AINS means a chatbot wearing a blazer.

There is a bigger risk. Angular Ventures, one of the more skeptical voices on this thesis, points out AINS is barely two years old as a concept. Nobody has a ten-year track record proving these margins survive a downturn, a regulatory crackdown, or a well-funded incumbent response. Deloitte did not go quietly after Saper's original "Death of Deloitte" piece; it sent lawyers, and its revenue has grown every year since. Big incumbents have capital, client relationships, and lobbying power that a bootstrapped owner-operator does not.

And the single point of accountability that makes AINS attractive to clients ("you buy the result, we own the delivery") also means you now own every failure mode that used to be diffused across a team of humans with individual liability and licenses. A lawyer who blows a filing deadline is one person's mistake. An AI-native law firm that blows a filing deadline because a model hallucinated a citation, as several firms have already discovered in federal court, is the firm's mistake, in public, with sanctions attached. Own the delivery infrastructure. Also own what breaks in it. Run the casualty drill before you need it, not after.

FAQ

Q: Is AINS just a rebrand of "services-as-software"? Mostly no, and the distinction matters. "Services-as-software" describes the workflow tooling layer, software that makes a human services team faster. AINS describes the business model: a company that sells the completed outcome, is the single point of accountability for that outcome, and prices accordingly. TripleDart is not selling you SEO software. It is selling you pipeline, at a flat fee, and it eats the labor cost of producing it.

Q: Do I need venture funding to build an AINS business? No, and TripleDart is the proof. $7M ARR, 50% EBIT margin, zero outside capital. What you need is not a term sheet. You need the discipline to codify your delivery process into a system before you try to price it as an outcome. Funding buys speed. It does not buy the discipline.

Q: What size business can actually make this pivot? Owner-operators between roughly $500K and $5M in revenue are the sweet spot. Below that, you likely lack the client base to find the repeatable patterns (Crosby's Ryan Daniels put the pattern-recognition threshold around 80 clients). Above $5M, the pivot gets harder because you have more legacy process, more people whose jobs are the manual work you are trying to automate, and more organizational resistance to compartmentalizing what they do.

Q: What is the first move if I want to try this? Write down your delivery process end to end, the way a Navy watchstander writes a standing order. Then split it into what a machine can run today and what still requires human judgment. Automate the first bucket. That is the whole first move. Everything else, outcome pricing, the multiple, the exit, follows from having an actual system instead of tribal knowledge in one person's head.

Q: What is the biggest way this fails? Leading with "cheaper" instead of "better," and skipping the accountability planning for when the AI-native delivery breaks. Both are avoidable. Neither is optional to think through before you launch.

Doctrine Connection

Systems beat slogans. AINS is not a slogan, and it is not a funding round. It is a system: codify the delivery, price the outcome, own the infrastructure instead of renting it from whatever platform currently takes a cut of your labor. Emergence Capital's summit at Crosby made headlines because a VC put a name on something. TripleDart made the actual case because it did the work without anyone's permission or capital. That is the Sovereignty Stack in one sentence: the operator who owns the system owns the multiple. The operator who rents it owns a job.

*Jeff Barnes has no personal position in any company, fund, or platform named in this article. demg.ai has no current commercial relationship with any party mentioned. demg.ai provides marketing education and operator strategy, not investment advice. Past performance does not guarantee future results.*