Tool review · Legal
Harvey
The enterprise legal assistant — firm workflows, not a gavel on ChatGPT.
The Aiacent take
Harvey is the legal AI we would shortlist when a firm already lost the ChatGPT argument with IT and still needs generation on matter files. Assistants, workflows, a knowledge layer. That is a platform. It is not 'ChatGPT for lawyers' with a better landing page.
It leads our legal ranking because 'best AI tools for lawyers' at a firm is usually this buying meeting. CoCounsel wins shops already in Thomson Reuters. Spellbook wins the contract in Word. Claude is what a careful solo actually opens — with a skip-if you should not ignore.
This is software, not counsel. Privileged files in a tool you have not contracted for are a malpractice story, not a productivity win.
Most 'we need Harvey' tickets from a five-lawyer shop are a Spellbook or Claude ticket with extra keynote.
Pros
- The legal-AI name large firms actually put through security
- Workflows for research, drafting, and review that are not a naked chat window
- Built for privilege, audit, and a knowledge base you will govern
Cons
- You will implement it like software. There is no honest $20 Harvey
- Overkill for a solo who needed a redline in Word
- A model that sounds like a partner is still not a partner. You sign the brief
Who it's for
- AmLaw-shaped firms and in-house teams with a security review
- Practices that will train the tool on their own work product
- Buyers whose ChatGPT ban is the actual ticket
Who should skip it
- Solos and small shops that needed Word markups (Spellbook)
- Researchers already paying for Westlaw / CoCounsel
- Anyone who thought a chatbot could replace a lawyer
What matters in the product
Matter workflows
The product. Research, draft, review on a file you still own. Empty knowledge is expensive autocomplete.
Firm controls
Why it wins RFPs. If you cannot answer where the document went, you should not have uploaded it.
Verdict
The enterprise legal desk. Shortlist Harvey when the firm will implement it. Do not shortlist it as a smarter ChatGPT, and do not treat the output as advice.
See Harvey