Harvey AI

Harvey AI is the most talked-about legal AI platform on the market, and unusually for a hyped product, the substance mostly matches the noise. It gives law firms a secure environment for legal research, drafting, contract analysis, and large-scale due diligence, with models tuned on legal material rather than a thin wrapper over a general-purpose chatbot. Its customer list — Allen & Overy, PwC Legal, and a growing roster of AmLaw firms — tells you exactly who it is built for.

What we like

The diligence workflows are the standout. Pointing Harvey at a data room of hundreds of contracts and getting structured, citable answers about change-of-control clauses or termination rights is a genuine time-saver measured in associate-days, not minutes. Research answers come with citations you can verify, and the workflow builder lets firms encode repeatable processes instead of reinventing prompts. Security and confidentiality controls are strong enough that risk-averse firm committees have actually signed off, which is its own endorsement.

Where it falls short

There is no public pricing, no trial, and no path for a solo lawyer or three-person firm to buy it. The enterprise-only model means most of the legal market simply cannot evaluate it. Like every LLM product, it still produces confident errors on edge cases, so the review burden never disappears — it shifts. And because Harvey is evolving fast, features you saw in a demo may differ from what your firm actually deploys.

Verdict

If you run knowledge management or innovation at a large firm, Harvey belongs on your shortlist next to CoCounsel — pilot both on a real matter. If you are a small firm or solo, look at Spellbook instead; Harvey is not built for you, and it does not pretend to be.