Harvey AI
Overview
Harvey AI is an enterprise AI platform for law firms and legal departments. It uses LLMs tuned for legal work to support research, contract drafting and review, due diligence, memo writing, and case law analysis. More than 200,000 legal professionals across 2,400+ legal organizations in 70+ countries now use it, including 75+ Am Law 100 firms.
Key Features
Assistant
Conversational AI for legal questions, drafting, and research, with prompts and templates tuned for legal tasks.
Vault (Document Analysis)
Holds up to 100,000 documents per vault for bulk analysis — powerful for M&A due diligence, litigation discovery, and large-scale contract review, including review against your firm's own precedent bank. Harvey reports 96% accuracy on key-term extraction, and Vault connects to iManage, SharePoint, and Google Drive so documents come straight from existing systems.
Agents (Harvey II)
Agents that carry legal work through end to end. With Harvey II, announced 18 August 2026, agents open inside a matter or project with its documents, parties, tasks, permissions, and history already in place, so lawyers no longer re-explain context on every task. Harvey also learns formatting, citation style, and writing preferences from your edits and carries that memory across Harvey, Word, and Outlook.
Workflows
Configurable multi-step processes — automate contract review → risk analysis → amendments flows.
Underlying Models (Harvey Tenet + third-party models)
Harvey now ships its own model, Harvey Tenet — its first model post-trained end to end for legal reasoning, which it says matches the strongest general-purpose models at an open-source cost. Third-party models remain selectable alongside it: GPT-5.6, Claude Opus 5, Claude Sonnet 5, and Fable 5 are all live in Harvey.
Collaboration Tools
Team projects with permission controls — ensure only authorized personnel access sensitive matters.
Supported Tasks
- Legal research
- Contract review
- Due diligence
- Memo writing
- Case law analysis
- Regulatory compliance checks
Security
Strict legal-grade security:
- SOC 2 Type II
- Confidentiality
- Data isolation
Pricing
| Item | Cost |
|---|---|
| Per seat | ~$1,200/month (higher with scale and modules) |
| Minimum seats | 20 seats |
| LexisNexis integration | +$400-$600/lawyer/month |
| Annual entry point | ~$360K |
Pricing is not published; the working benchmark is roughly $1,200 per seat per month with a 20-seat minimum, putting the annual entry point near $360K. Sales cycles frequently run six months or longer, which in practice restricts Harvey to large firms and big in-house teams.
Typical Use Cases
- Big-law M&A due diligence
- Litigation discovery (mass document review)
- Contract review acceleration
- Legal memos
- Case law research
Strengths
- ✅ Vault — 100K documents per vault, 96% key-term extraction accuracy
- ✅ Harvey II agents inherit matter context and remember how you work
- ✅ Harvey Tenet in-house legal model alongside GPT-5.6 / Claude Opus 5
- ✅ Workflows for multi-step automation
- ✅ Adopted at 75+ Am Law 100 firms, 2,400+ legal organizations
- ✅ Strict security
Weaknesses
- ❌ Very expensive (~$1,200/seat/month, 20-seat minimum)
- ❌ LexisNexis integration costs an extra $400-$600 per lawyer per month
- ❌ For large/mid-size law firms only
- ❌ Not for solos or small firms
- ❌ Limited Japanese legal support
- ❌ Opaque pricing and long sales cycles
Official Resources
- Website: https://www.harvey.ai