Intapp

What it is

Intapp sells the layer that decides whether a firm is allowed to take a piece of work at all, and then governs who inside the firm gets to see it. Conflicts clearance, new-business intake, outside counsel guideline terms, ethical walls, employee compliance attestations, time capture and billing — plus DealCloud, the deal and relationship CRM used on the private capital side. The company grew out of an integration business launched in 2002 and has been public on NASDAQ since 2021 under INTA. It counts 96 of the Am Law 100, 16 of the top 20 global accounting firms, and more than 1,700 private capital and investment firms as clients.

This is the part of the legal stack the ooligo catalog had not covered. Harvey, Legora and Spellbook work at the practice layer, on the work product. Intapp sits underneath, at the firm-governance layer that decides which lawyers are permitted to touch which matter — and it spent 2026 converting that position into an AI control point.

Three things moved. On February 23, 2026, Intapp and Harvey announced that Intapp Walls policies sync into Harvey and enforce across Assistant, Vault and shared workspaces, with access blocked when authorization cannot be confirmed and every enforcement decision logged on both sides; that integration reached GA for cloud Walls in July 2026. The day after, on February 24, Intapp announced a collaboration with Anthropic to build industry-specific agents powered by Claude inside Intapp’s governed environment. Then on July 15, 2026, Celeste — the agentic layer Intapp calls “firm AI” — reached GA after a February preview, with BakerHostetler named on intake and conflicts and the private equity firm Hg named as an early adopter.

Why it shows up in Legal Ops stacks

• Conflicts checking and intake are one system, not two. Intapp Conflicts runs the search and the disposition; Intapp Intake carries the risk questionnaire, the approval chain and the AML and KYC steps. A firm clearing thousands of new matters a year wants the clearance record and the intake decision in the same audit trail, because that is the artifact a client’s outside counsel audit asks for.
• Walls is where AI governance actually gets enforced. Policies defined once in Intapp Walls apply across Harvey, Microsoft Copilot and Celeste. The alternative is re-implementing lateral-hire screens, matter-level restrictions and MNPI controls separately inside every AI tool the firm buys, then defending that patchwork to a client.
• Celeste targets the business of the firm rather than the brief. Playbook-driven agents run deal screening, conflicts clearance preparation, business development research, pricing, and the lateral hiring lifecycle from candidate research through onboarding — the work that sits between matter management and firm strategy.
• The financials say firms are expanding, not just renewing. Cloud ARR was $459.3 million at March 31, 2026, up 31% year over year and 82% of total ARR. SaaS revenue in fiscal Q3 was $107.9 million, up 27%. Trailing twelve-month cloud net revenue retention was 123%.

Pricing reality

Nothing is published. There is no pricing page and no per-module rate card — every deal is quoted through sales.

The disclosed customer bands are the honest anchor. As of March 31, 2026, Intapp reported more than 1,375 clients above $50,000 of ARR, and 858 of those above $100,000. Read the first number as the effective entry point: a single module at a mid-size firm is a five-figure annual commitment, and about six in ten of the firms on the books have crossed six figures, which is where two or three modules land. Third-party buyer data circulating for the platform puts the full range at approximately $85,000 to $1.4 million a year depending on module count and firm size. Treat that as an estimate rather than a quote, but its shape matches the disclosed bands.

Modules move the number far more than headcount does. Ask for Conflicts, Intake, Walls, Terms and Time as separate line items, and get Celeste quoted apart from the modules it reads.

Best for

Risk partners, general counsel’s offices and legal operations leaders at firms above roughly 150 lawyers, where a conflicts miss is an existential event rather than an embarrassment, and where lateral hiring volume makes screen administration a standing job rather than an occasional one. It is the strongest pick for a firm deploying practice-layer AI at scale and now being asked by clients how access is controlled, because Walls answers that question with an audit log instead of a policy document.

Do not buy it for a 30-lawyer firm. At that size, conflicts clearance is a searchable client database plus a partner who knows the book, and the implied six-figure platform cost buys a risk professional instead. Do not buy Celeste expecting it to replace practice-layer AI either — it runs the business of the firm, not the work product, and firms that conflate the two end up paying for both and using neither well.

Versus the alternatives

Aderant, a Roper Technologies business, is the credible number two, and it bundles the problem differently: Conflicts and Risk sits inside Aderant Expert practice management, so the conflicts record lives next to the billing record. On May 11, 2026, Aderant launched its Agent Center on the Stridyn platform with Collections, Appeals and Talent Evaluation agents, and it runs its own Harvey partnership. Pick Aderant when the firm is already on Expert and the committee wants one vendor across finance and risk. Pick Intapp when risk and compliance is the load-bearing requirement and you accept running it alongside a separate practice management system.

iManage Security Policy Manager is the alternative for walls specifically, and it wins on scope. If the firm’s content lives in iManage repositories, SPM enforces need-to-know at client, matter, department and location level against that repository with less integration work than a platform purchase. Pick it when document security inside the DMS is the entire problem. Pick Intapp Walls when the walls have to reach past the DMS into AI tools, time entry and CRM.

Legora is the fastest-growing entrant in legal AI, and it competes for the same budget line rather than the same job — $100 million ARR in April 2026, up from $3 million at the end of 2024, on a $600 million Series D at a $5.6 billion post-money valuation announced April 30, 2026. Pick Legora when the partnership’s complaint is that associates are slow, not that intake leaks. These two are not substitutes: a firm that buys Legora without a walls policy has bought its next audit finding, and the difference between legal AI and legaltech is exactly this split between doing the work and governing it.

If none of the three fit, the problem is that nobody owns risk as a function. Software enforces a policy; it does not write one. A firm with no documented lateral screen procedure gets a faster version of the same ambiguity.

Watch-outs

• Celeste has been GA for under a month and has two named references. BakerHostetler on intake and conflicts, Hg on the private capital side. Everything else about how it behaves against your firm’s data volume is vendor claim. Guard: scope the first Celeste purchase to a single playbook with a measurable clock — lateral hire clearance turnaround is the cleanest — capture the baseline before go-live, and negotiate a 12-month exit on that module rather than signing it into a three-year platform term.
• Walls for AI enforces the policies it holds, not the ones you intended. The Harvey sync propagates existing Intapp Walls policies. A screen that was never entered, or a matter mis-coded at intake, becomes confident enforcement of the wrong rule across every connected AI tool at once. Guard: audit policy data before enabling the sync — pull 20 active screens and verify each against its source memo — and make the enforcement log a monthly review with a named owner, not a feature you switch on and forget.
• Every price is a quote, and auto-renewal is the default posture. Third-party buyer reports describe annual contracts with auto-renew clauses that are hard to strike. Guard: negotiate the renewal notice window and the uplift cap in the first order form, not at renewal, insist on per-module line items, and diary the notice date the day you sign.
• The AI is retrofitted onto a 2002-vintage compliance platform, and the frontier model belongs to someone else. Claude comes from the Anthropic collaboration and the practice-layer intelligence comes from Harvey; Intapp’s own contribution is governance, playbooks and the firm’s data. That is a defensible position, but it means the roadmap depends on two partnerships. Guard: ask which Celeste capabilities are GA in your tenant on the contract start date and get the list in writing. Price the deal on the compliance modules and treat the agents as upside.
• Deployment runs in quarters and the value case starts after migration, not after signature. Third-party implementation reports put a single module at roughly 8 to 20 weeks and a multi-office rollout with practice management and DMS integrations at 3 to 9 months — estimates, not vendor commitments. Guard: stage the purchase, Conflicts and Intake first and Walls once the matter data is clean, and tie payment milestones to go-live per module.