# Harvey — Proof, Not Just Trust: How Harvey is Building Reviews and Verification

- Company: Harvey (harvey.ai)
- Announced: 2026-11-19T18:00:00+00:00
- Category: capability-change
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://www.harvey.ai/blog/how-harvey-is-building-reviews-and-verification
- Record: https://forck.live/items/38787-proof-not-just-trust-how-harvey-is-building-reviews-and-verification
- Subject: Harvey

Harvey is building verification into its legal AI, saying lawyers who sign their names to work need proof of how an answer was produced, not just trust. The company argues that a diligence memo can read cleanly and still be wrong because the agent never opened a side letter or missed an amendment, and that checking every citation by hand does not scale at the speed agents now work. Harvey says it checks what can be checked automatically: citations, quotes and pin cites, building on its 2024 BigLaw Bench sourcing evaluation, agentic search in legal research, Harvey Graph and its open-source Legal Agent Benchmark. Cite Check, coming to Early Access later this year, will find every citation in an uploaded brief and check each quote and pin cite against the full opinion in US Case Law, marking citations verified, needs review or issue found and offering fixes as tracked changes. For judgment calls a machine cannot check, Harvey points to thinking states that show which files it is searching and reading, and says answers will carry a short scope note naming the documents read in full and any files it could not open. Harvey says early design partners trusted a plain account of coverage more than a coverage percentage, and it is exploring letting Harvey pause on a judgment call, ask the lawyer and record the decision.

## Evidence

Verbatim from https://www.harvey.ai/blog/how-harvey-is-building-reviews-and-verification:

> Lawyers sign their name to their work. AI should give them clear proof, not just ask for trust.
> 
> A lawyer signs their name to every piece of work that leaves their desk, even when Harvey wrote the first draft. As Harvey takes on more of that work, from reviewing tens of thousands of documents in a single table to working through an entire data room for a diligence memo, verification becomes the bottleneck.
> 
> In most fields, the output tells you most of what you need to know. An engineer can run AI-written code. A designer can judge a screen on sight. But a diligence memo can read cleanly and still be wrong because the agent never opened a side letter or missed an amendment. The only way to know is to see how the answer was produced: what was read, what was skipped, and what it relied on. In legal, the work itself is part of the output.
> 
> Today that check is the lawyer: a partner reading a draft line by line, or an associate clicking through every citation. That doesn’t scale at the speed agents now work.
> 
> So we think about solving verification for legal AI across three dimensions:
> 
> What Harvey checks on its own: citations, quotes, and pin cites.

---

Record: https://forck.live/items/38787-proof-not-just-trust-how-harvey-is-building-reviews-and-verification
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md
