Centari

What it is

Centari is a deal intelligence platform for transactional law firms and investment funds. It reads a firm’s closed-deal documents (purchase agreements, credit facilities, partnership agreements, side letters) and turns them into structured, citation-backed data. Deal teams can then search, compare, and chart that data across hundreds of transactions. It is not a drafting assistant, and it does not write the markup for you. Its job is to tell you what your firm has actually agreed to before, so the next negotiation starts from evidence rather than memory.

The core is what Centari calls a patent-pending Deal Reasoning Engine. It combines LLM orchestration, proprietary document processing, and calibration by in-house legal experts to extract terms the way a transactional lawyer reads them: following defined terms, chasing cross-references, and tracking which document modifies which. The 2026 releases built on that engine:

• Intelligence and Views (March 10, 2026): natural-language market analysis and clause search over the firm’s deal data, plus dashboards for internal teams.
• Amendment Awareness and Deal Maps (June 2026): deal data updates when an amendment is uploaded, so it shows the operative terms rather than the original language. Deal Maps draws a graph of how the documents in a transaction incorporate, modify, satisfy, or supersede one another.
• External Views (July 8, 2026): white-labeled client dashboards with per-person permissions on charts, tables, benchmarks, and supporting documents. They can show client-specific data or anonymized aggregate market data.
• Precedent Intelligence (August 13, 2026): Precedent Search retrieves a provision by describing what it does rather than guessing its wording. Precedent Capture lets deal teams adjust or add citations, with every edit tracked. Multi-Doc Compare lines up a provision side by side across deals, filtered by deal type or counterparty.

CEO Kevin Walker founded Centari in 2023 in New York. He is a former Paul Hastings M&A lawyer and deputy general counsel, and the team includes lawyers from Kirkland & Ellis and Davis Polk. On September 18, 2025 it announced $14M in total funding, including an $8.4M Series A led by Sentinel Global, with South Park Commons and GTMfund participating. Named customers include Ropes & Gray, Willkie, Fried Frank, Wilson Sonsini, Goodwin, and Gilbert + Tobin. It also works with Sentinel Capital Partners on the fund side, and it operates in the US, UK, and Australia. It holds SOC 2 Type II and ISO 27001 attestations, offers single-tenant deployment, and states that customer data is never used for training or commingled.

Why legal ops teams look at it

The question it answers is “how do we get market-standard terms out of our own closed deals instead of guessing at fallback positions?” At most firms, that knowledge sits in a senior partner’s head or in a closing set nobody has opened since signing. Public benchmarks such as deal-terms studies describe the market. They do not describe what your firm conceded to this sponsor on its last four deals.

Specific use case: a private equity M&A group runs Centari over five years of its own closed deals. When a sponsor client asks “is a 12-month survival period on general reps still market for us?”, the associate uses Multi-Doc Compare, filtered to that sponsor and deal size, and answers with cited precedent in minutes. The practice’s KM lead then publishes an anonymized External View of indemnity-cap trends to the client. That turns the deal archive into a business development asset.

Pricing reality

Centari does not publish prices. It sells through enterprise contracts, and Legaltech Hub lists its target as firms with 500+ lawyers. That puts it in the same buying motion as DraftWise and Kira, where AmLaw-scale deployments are reported in the high-six to low-seven figures a year. Treat that as a directional band from comparable tools, not a Centari quote. Two cost drivers matter more than the per-seat line: how many closed deals you load in the first year, and whether you pay for single-tenant hosting. Budget internal time as well. Precedent Capture only pays off when associates or KM staff correct citations, and that work is not free.

Best for

Knowledge management leaders, legal ops directors, and practice group heads in M&A, private equity, finance, and fund formation practices at firms with a large archive of closed deals. It fits teams that want that archive to answer “what is market for us” questions and to feed client-facing market reports. Fund-side deal and legal teams with a large portfolio of credit and governance documents are a second audience.

Alternatives and when to pick them

• Kira (Litera): the long-standing leader in M&A clause extraction and due diligence. Pick it when the job is reviewing a target’s data room at volume, not mining your own precedent, and when you already buy the Litera bundle.
• Harvey: the legal AI assistant with the largest firm footprint. Pick it when you want one broad tool for drafting, research, and review. It reads documents well, but it does not keep your deal archive as a structured, amendment-aware dataset.
• Legora: the fastest-growing entrant in legal AI. Pick it for a collaborative workspace with tabular review across a document set at a lower entry price than Harvey. The same caveat on structured deal history applies.
• DraftWise: the closest head-to-head. Pick it when drafting in Word from precedent is the priority, since DraftWise lives in the Word add-in and indexes iManage and NetDocuments directly. Pick Centari when the priority is analysis across deals, amendment tracking, and client dashboards.
• Hebbia: pick it on the fund side when the corpus is mostly financial and diligence material rather than executed legal agreements.

Watch-outs

• No documented DMS integration or API. None of Centari’s 2026 announcements names an iManage or NetDocuments connector, a Word add-in, or a public API. Guard: before signing, get the ingestion path in writing (connector, export, or bulk upload), who maintains it, and how new closings reach the platform without a manual step.
• Output is only as good as the calibration. The data-layer approach depends on correct extraction, and the vendor itself says firms worry about putting their name on client portals built on unreliable AI. Guard: in the pilot, have two associates spot-check extracted terms on 20 deals they closed. Do not publish an External View until the error rate on those deals is one you would defend to a client.
• Client confidentiality in shared views. External Views can expose aggregate market data drawn from other clients’ deals. Guard: have the general counsel’s office approve the anonymization threshold (a minimum deal count per data point) and review each client-facing view before it goes live.
• Thin independent review record. Centari has no meaningful G2 or Capterra presence, and nearly all public coverage is vendor press. Guard: ask for reference calls with two firms of your size, one of them outside the launch customers named in press releases.

Background: contract data extraction and legal knowledge management.