Intapp’s Q4 FY26 earnings call introduced Celeste, its new “agentic platform for Firm AI.” Strong quarter, confident presentation, a genuinely impressive slide deck … until you get to the slide showing their “comprehensive platform built for professional firms.”

Look past the AI framing at what Celeste is actually doing – sitting on top of Intapp’s own product line — DealCloud for growth, Intake/Conflicts/Terms for compliance, Time/Billstream for profitability — and also plugging into Microsoft Copilot, Claude, and Harvey to cover whatever those don’t.

Celeste’s own positioning describes itself as something that “enhances rather than replaces” the AI tools firms are already stitching together.

That's not a product decision. That's an admission.

The tell shows up at the point of use

A partner at a patent and IP law firm put it plainly: his marketing team is using DealCloud, and it isn't working for them.

That's not a training issue. DealCloud was built for deal pipelines — sourcing, diligence, close. It's genuinely strong at that job, which is why private equity and investment banking firms run on it.

But a patent and IP practice doesn't move prospects through a pipeline with a close date. It runs on long-standing relationships, referral networks, and knowing who's connected to whom over years, not a deal stage.

Ask a system built around pipeline stages to do relationship-driven marketing, and it will fight the team using it every day, because the underlying data model was never designed to answer the questions marketing actually has: who do we know, how are they connected, what's the history, who owns this relationship.

Intapp's answer to that mismatch isn't to fix the data model. It's to include DealCloud as one pillar in a larger stack, then build an AI layer to reason across all the pillars at once. Celeste's own "Context Engine" — combining firm context, ontology, and memory across systems — exists specifically because the underlying systems don't share a data model in the first place. It's an AI product whose job is to compensate for the seams between the other products.

Why this breaks cross-functionally, not just for marketing

The marketing team's DealCloud problem doesn't stay contained to marketing.

If relationship data lives in a deal-pipeline tool, and client history lives somewhere else, and time/billing lives in a third system, then every cross-functional question — has anyone here worked with this contact before, is there a conflict, who should make the introduction — requires reconciling across systems that weren't built to talk to each other natively.

Celeste is Intapp's bet that AI can do that reconciliation well enough, fast enough, to make the seams invisible. That's a real engineering effort. It's also a workaround for a structural decision made years earlier: build separate best-of-breed products, sell them separately, then bolt them together later.

The alternative isn't a smarter seam. It's no seam.

Venntive's architecture starts from the opposite premise — one platform, one UI, one database.

Relationship data, marketing activity, deal or engagement history, communications, time — all of it lives in the same system, viewed through the same record.

There's no reconciliation step, because there was never a second data model to reconcile against.

When a relationship-driven firm asks "who do we know at this company," the answer isn't assembled by an AI layer inferring across five products in real time — it's just there, because it was always one system.

This is the actual argument behind Venntive's "Power of One" positioning, and it's worth being precise about what it means — it isn't marketing shorthand for "simple."

It's a structural claim about where the data lives and how many systems have to agree with each other before a person doing their job gets a straight answer. Twenty-five years of building revenue systems for B2B firms comes down to a consistent finding — the firms that struggle aren't the ones with the wrong software feature, they're the ones running five tools that were each right for someone else's business and never designed to be right together.

Who this is for — and who it isn't

If a firm's actual work is deal-flow — sourcing, diligence, a defined close — DealCloud earns that use case honestly, and Celeste's bet on AI-mediated coordination across a modular stack may pay off for firms already committed to that architecture.

But for relationship-driven professional services firms — law, accounting, consulting practices whose business is ongoing client relationships rather than transactions with a close date — a deal-pipeline tool with an AI layer bolted on top to compensate isn't the fix. The fix is a system that was built around relationships in the first place.

Book a Quick Q&A to see whether your firm's structure fits a unified system, or whether you're solving the wrong layer of the problem.