🇺🇸 United States · Harvey — AI for Legal Professionals
Status: 🟩 COMPLETE 🟦 LIVING Section: 10 — AI and LLMs
| Vendor | Harvey AI |
| Country/origin | 🇺🇸 United States (San Francisco) |
| Recommended for AUS? | ✅ Yes — US company; enterprise legal data handling; used by major Australian law firms |
| Privacy summary | Enterprise-grade; no training on client data; SOC 2 Type II; data residency options; works with Australian law firms under enterprise agreements |
| Free tier | No — enterprise only |
| Paid tiers | Enterprise contracts only (pricing not public; typically per-seat or usage-based) |
| First released | 2022 (stealth); public awareness 2023; rapid adoption 2024 |
| Last reviewed | June 2026 |
| Official site | https://harvey.ai |
What it is
Harvey is an AI platform built specifically for law firms and legal teams — designed to handle legal work that would be dangerous or inaccurate if done by a general-purpose AI like ChatGPT.
Founded by Gabriel Pereyra (ex-Google DeepMind) and Winston Weinberg (ex-Goldman Sachs lawyer), Harvey raised enormous backing from OpenAI and major law firms, and quickly became the dominant AI platform inside large law firms globally.
What Harvey does that general AI can’t do well:
- Legal documents are highly technical, jurisdiction-specific, and where mistakes have serious consequences. A general AI might hallucinate a case citation, misstate a legal standard, or give advice appropriate in the US but wrong in Australia. Harvey is trained and calibrated specifically on legal content with appropriate guardrails.
Key capabilities:
- Contract drafting and review: Draft standard legal agreements; identify unusual or missing clauses; compare contract terms against a client’s playbook
- Contract analysis at scale: Review thousands of contracts for specific provisions (e.g., change-of-control clauses, governing law, liability caps)
- Legal research: Research case law, statutes, and regulatory materials — with citations that are verified (not hallucinated)
- Due diligence: Accelerate M&A (mergers and acquisitions) due diligence by reviewing large document sets
- Regulatory analysis: Analyse regulatory filings, government guidance, and compliance documents
- Matter summarisation: Summarise large case files, discovery productions, and document sets
- Litigation support: Draft briefs, summarise depositions, identify key documents
Who uses Harvey
Harvey is used by major law firms globally:
- In Australia: Clayton Utz, Herbert Smith Freehills, Allens and other top-tier firms
- Globally: Allen & Overy (now A&O Shearman), Latham & Watkins, PwC Legal, many others
- In-house legal teams at major corporations
Why law firms need specialised AI (not just ChatGPT)
-
Hallucinated citations are career-ending. Multiple cases of lawyers submitting AI-generated briefs with fake case citations (most famously the Mata v. Avianca case, 2023) resulted in sanctions and professional discipline. Harvey verifies citations against real legal databases.
-
Jurisdiction matters. Australian contract law differs from US contract law; Australian employment law differs from UK employment law. A model trained on primarily US legal content gives advice that may be simply wrong in Australia. Harvey is trained with jurisdiction awareness.
-
Confidentiality. Legal work is bound by legal professional privilege and confidentiality obligations. Harvey’s enterprise terms ensure client matter content is not used for training and is not accessible to other law firms.
-
Professional responsibility. Lawyers have duties to supervise work product — including AI-generated work. Harvey is designed with appropriate human review workflows rather than end-to-end automation.
Australian legal context
- Australian law firms are among the early global adopters of Harvey
- The Law Society of NSW, Law Council of Australia, and various bar associations have published guidance on AI use in legal practice (as of 2024–2026)
- The general position: AI can assist in legal work but lawyers retain professional responsibility; AI output must be reviewed and supervised
- Harvey’s Australian-specific capabilities include: Australian case law databases, Australian regulatory framework content, Australian contract conventions
How to access
Harvey is enterprise-only — there’s no self-serve sign-up. Access is through a law firm that has subscribed, or by contacting Harvey’s enterprise sales team. Individual lawyers outside of Harvey-subscribing firms cannot easily access it.
For Australian law firms: contact harvey.ai/enterprise for a demonstration and pricing.
Competitors in legal AI
| Tool | Country | Notable |
|---|---|---|
| Harvey | 🇺🇸 | Largest; deepest law firm adoption |
| Spellbook | 🇨🇦 | Contract drafting; works inside Microsoft Word |
| CoCounsel (Thomson Reuters) | 🇺🇸 | Integrated with Westlaw legal research database |
| Lexis+ AI | 🇺🇸 | Integrated with LexisNexis database |
| LawGeex | 🇮🇱 | Contract review specialist |
| Paxton AI | 🇺🇸 | US legal research focus |
| M-Files + AI | 🇫🇮 | Document management with AI for legal |
Gotchas
- Not accessible to individual lawyers or members of the public. Harvey requires an enterprise subscription through a law firm or legal department.
- AI legal tools are not DIY legal advice. If you’re an individual needing legal advice, Harvey won’t help you directly. You need a lawyer (who may use Harvey as part of their work).
- Australian case law integration is still developing. Some Australian-specific legal databases are less deeply integrated than major US/UK sources. Australian lawyers should verify coverage.
- Human review is still required. Harvey dramatically accelerates legal work but doesn’t replace lawyer judgment. It’s a power tool for lawyers, not a replacement for lawyers.
- Pricing is significant. Enterprise AI legal tools are expensive. Smaller Australian law firms may find the cost prohibitive. Mid-market alternatives (Spellbook, built into Microsoft Word) may be more accessible.
See also
- spellbook — contract AI inside Microsoft Word; more accessible
- notebooklm — free document analysis for non-legal research
- deep-research-mode — AI research tools for non-legal work
- ai-document-generation — AI for general document creation
Sources
- Harvey AI official site: harvey.ai
- A&O Shearman × Harvey partnership announcement (2023)
- Clayton Utz AI adoption announcement
- The American Lawyer, Legal Futures coverage of Harvey (2023–2026)
- Law Society of NSW AI guidance (2024)
- Law Council of Australia AI position statement (2024)
- Mata v. Avianca (2023) — the landmark AI legal citation hallucination case