Harvey

Score9 out of 10

1 Reviews and Ratings

What is Harvey?

Harvey is an enterprise-grade generative artificial intelligence platform engineered specifically for legal research, contract analysis, and regulatory workflows. The system leverages custom-trained Large Language Models (LLMs) to assist corporate legal departments, compliance officers, and law firms in analyzing vast corpora of unstructured legal documents and generating complex legal drafts.

Key Capabilities

Ambient Legal Research (Harvey Assistant): Employs advanced Natural Language Processing (NLP) to respond to complex legal queries, analyze case law, and draft legal arguments, delivering source-verified citations to allow legal professionals to validate model outputs.
Large-Scale Document Extraction (Vault): Automates the analysis of thousands of documents concurrently (such as during merger and acquisition due diligence reviews), allowing users to extract specific clauses, analyze structural risks, and isolate data points across vast datasets.
Workflow Automation (Harvey Agents): Deploys specialized software agents designed to orchestrate and execute multi-step legal workflows, streamlining complex administrative and research tasks.
Deep Productivity Integration: Integrates with established enterprise communication platforms and Document Management Systems (DMS)—including Microsoft 365, iManage, NetDocuments, and SharePoint—via native Application Programming Interfaces (APIs).

Audience & Use Cases

Audience: Corporate legal counsel, litigation attorneys, compliance officers, and law firm administrators.
Use Case: Automating contract lifecycle analysis, accelerating due diligence reviews, drafting legal briefs, and researching jurisdictional precedents.

Technical Specifications

Compliance & Security Standards: SOC 2 Type II, ISO 27001, and ISO 27701 certified; compliant with GDPR and CCPA data privacy guidelines.
Data Sovereignty: Employs AES-256 encryption at rest and TLS 1.2+ encryption in transit, supporting Bring Your Own Key (BYOK) options and guaranteeing that customer inputs are not used to train public models.