Overview
Harvey is an AI platform built specifically for legal professionals. It assists with contract analysis, legal research, due diligence, document drafting, and knowledge management. Used by top law firms and in-house legal teams worldwide, Harvey understands legal context, citations, and professional standards.
Our take
Harvey is AI built specifically for legal and professional services, sold through enterprise deployment. Domain specialisation matters more in law than almost anywhere else, because the failure mode — a confidently fabricated citation in a filing — has produced sanctioned lawyers. Grounding in real legal sources with verifiable citations is the property that counts.
Worth it if
Law firms and legal departments with the scale to justify enterprise deployment and training.
Look elsewhere if
Solo practitioners and small firms — the deployment model assumes an organisation.
✨ Features
Legal Research
AI-powered research across case law, statutes, and legal databases.
Contract Analysis
Reviews and analyzes contracts, identifying risks and key provisions.
Document Drafting
Generates legal documents, memos, and briefs based on firm standards.
Knowledge Vault
Centralizes firm knowledge and work product for AI-powered retrieval.
⚖️ Pros & Cons
Pros
Cons
💰 Pricing
Enterprise
Custom
- Full platform access
- Custom AI model training on firm data
- Enterprise security and SSO
- Dedicated implementation team
- Contact sales for pricing
Categories
❓ Frequently Asked Questions
Does specialisation actually prevent hallucinated citations?
It reduces them substantially when answers are retrieved from a real legal database and linked to the source, rather than generated from the model alone. It does not eliminate the need to verify. Every citation should resolve and say what the document claims before anything leaves the firm — that duty stays with the practitioner.
What does client confidentiality require from a vendor?
A contract excluding training on your content, a stated retention period, ideally zero-retention processing, a data processing agreement, and clarity on sub-processors and data location. Professional confidentiality obligations do not relax because a privacy page is reassuring, and the duty to supervise technology sits with the practitioner in most bars.