Intapp Isn't Selling AI. It's Selling the Rules Around It.

Intapp announced a new integration with OpenAI yesterday. DealCloud, Intapp's platform for financial services deal-making, now works as a plug-in inside ChatGPT for Financial Services. Deal teams at private equity firms and investment banks can query their own client records, deal history, and relationship maps directly from ChatGPT.

Most press releases about AI partnerships sound the same. But the one thing almost no one is talking about is the actual constraint these firms face. It isn't the AI model. It's the rules around what people are allowed to see.

If you're a lawyer at a large firm, you cannot look at every client file. If you're a banker at Goldman Sachs, you cannot cross certain internal information barriers. One missed conflict of interest can cost a firm millions in penalties and reputational damage. These are not IT problems. They are the reason these firms exist.

Intapp has spent 20 years building the compliance infrastructure that enforces those boundaries. The company started as legal software, formerly called LegalApp Holdings. Over two decades it accumulated the systems of record for 2,750 firms — including 96 of the 100 largest U.S. law firms and about 1,700 private capital and investment banking shops.

So the OpenAI announcement isn't really about Intapp getting access to better AI. It's about OpenAI going through Intapp to reach firm data at all.

This is the kind of moat you can't build by training a bigger model.

The market hasn't priced this in. Intapp trades at 5 times trailing revenue — a multiple that suggests investors see a growing SaaS company with an uncertain profitability path. They see negative GAAP earnings and a stock that is down about 17% year-to-date.

They're looking at the wrong problem.

Intapp's total cloud annual recurring revenue was $496 million as of June 2026, up 29% from a year earlier. Net revenue retention — the rate at which existing customers spend more over time — sits at 123%. That means for every dollar of recurring revenue a firm starts with, it gives back $1.23 a year later. Firms are expanding their Intapp contracts faster than competitors are eating them away.

But here's the more important number: cloud ARR now represents 84% of Intapp's total recurring revenue, up from 74% two years ago. The company is converting old on-premise installations to the cloud. And the cloud is where the AI works.

In February 2026, Intapp launched Celeste, what it calls an "agentic AI platform" for professional firms. Celeste doesn't just summarize documents. It runs firm workflows — screening deals, clearing conflicts, building pitch materials — on each firm's own data, under its own rules. In July, it reached general availability.Intapp has also connected Celeste to Claude from Anthropic and to Microsoft Copilot. The OpenAI plug-in is the third major model integration in seven months.

This pattern reveals something about Intapp's actual business model. The company isn't betting on any particular AI model. It's betting that whoever wins the AI model war still has to route through someone who understands professional compliance. Intapp wants to be that someone.

Most people think of AI as a capability problem. The bigger the model, the better the output. But professional services firms face a governance problem. The best model in the world is useless if you can't trust it with your firm's data. And you can't trust it until someone has built the rules into the pipe.

Intapp has been building that pipe since 2000.

There are reasons to be cautious. Intapp reported $40 million in GAAP operating losses for fiscal year 2026, wider than the $27 million loss the prior year. The losses come mostly from stock-based compensation, which is a real cost to shareholders through dilution, not just an accounting adjustment. Free cash flow of $136 million looks strong until you notice the company spent $275 million on share buybacks in the same year, cutting its cash balance from $313 million to $163 million.

The company is buying back stock while running GAAP losses and depleting cash. That's a signal that management believes the non-GAAP profitability story is the real one, and that the stock is undervalued even at these levels. It could also be a sign that the board wants to offset dilution from stock compensation by reducing share count, which is a common pattern in SaaS companies that pay employees in equity.

Neither interpretation makes Intapp obvious. What it does make clear is that the margin profile question is live.

The bigger risk is less about Intapp and more about its customers. Professional services firms have historically been slow software adopters. Lawyers and bankers use what works and switch only when forced. Intapp's 123% net revenue retention suggests the firms that do adopt Intapp stick with it and expand. But adoption velocity matters. Celeste reached general availability in July 2026. The OpenAI plug-in launched yesterday. It could be months before these products generate meaningful incremental revenue.

The OpenAI integration is also an implicit admission that Intapp needs the big platforms to reach end users. Firms want to use AI inside tools they already know — ChatGPT, Copilot, Claude — not inside Intapp. So Intapp is effectively licensing its data layer to the models it would most naturally compete against. The company says governance stays with Intapp regardless of where the query happens. That's true. But it also means Intapp is becoming a data pipe for companies whose primary business is selling AI access to the same customers.

Whether that's a durable position depends on something Intapp can't control: whether compliance gets commoditized. If OpenAI or Microsoft ships their own compliance layer built for financial services — rules for conflicts, information barriers, client confidentiality — then Intapp's governance moat starts to leak. That seems unlikely in the near term. Compliance isn't a feature you bolt on. It's something you earn by spending two decades building systems that lawyers and bankers actually trust. But it's not impossible.

Here's the test. Watch Intapp's cloud ARR growth over the next four quarters. The company just passed $500 million. If Celeste and the model integrations create real demand, cloud ARR should accelerate above the current 29% growth rate. If they're just feature additions on an existing platform, growth will hold steady — and the stock will remain valued as a good but not exceptional vertical SaaS business.

The market is pricing Intapp as a compliance software company that added AI. It may be worth more if it turns out to be an AI data layer that happens to sell to compliance-bound firms. The difference matters because the second story has much more room to grow.