Harvey
Legal AI for research, drafting, and document analysis
Harvey is a legal AI assistant built for research, drafting, and document analysis. Many large law firms and professional services teams deploy it to speed up the work attorneys do each day, from case law research to first-draft memos and contract review.
This review covers what Harvey does, its main features, how well it performs for legal work, how pricing works, who should use it, and how it compares to other options. The goal is a clear picture for firms weighing a legal AI tool.
What is Harvey?
Harvey is a legal AI assistant that helps attorneys research questions, draft documents, and analyze contracts inside one workspace. It is made by Harvey, a company founded in 2022 that focuses on AI for the legal profession. The product reached wider release around 2023 and has since been adopted by large law firms and professional services organizations.
The tool serves lawyers who handle high volumes of research and drafting. Instead of switching between a research database, a word processor, and a review tool, an attorney can ask Harvey a legal question, request a draft, or upload a contract for analysis. Harvey returns answers grounded in legal sources with citations the attorney can check.
Harvey targets the workflows of large firms and in-house legal teams. It is designed for matters where accuracy, confidentiality, and traceable sources matter. That focus shapes the product: it emphasizes grounded citations, enterprise data controls, and support for long, dense legal documents.
Key features
Harvey brings the core tasks of legal practice into one AI workspace. The main capabilities are:
- Legal research: query case law, statutes, and regulatory material and get answers with citations to the underlying authority.
- Document drafting: generate first drafts of memos, briefs, client letters, and other legal documents from a prompt or a set of facts.
- Contract analysis: review agreements, surface key clauses, flag issues, and compare terms across documents.
- Long-document handling: summarize and answer questions about lengthy filings, contracts, and records without manual page-by-page reading.
- Grounded citations: link answers back to source authority so attorneys can verify each point.
- Enterprise controls: manage access, data handling, and matter confidentiality across the firm.
The value of these features is in how they fit together. An attorney can research an issue, pull the findings into a draft, and then run analysis on a related contract without leaving the workspace. Citations run through the whole flow, so the output stays checkable rather than a black box.
How well does it work?
Harvey performs well on the research and drafting tasks it was built for. Its strength is depth in legal content: it handles multi-step legal questions, reads long documents, and returns answers tied to source authority. For a large firm processing many matters, that grounding is the feature that separates a legal AI tool from a general chatbot.
Strengths
- Grounded answers with citations that attorneys can verify against source law.
- Strong handling of long contracts, filings, and records that would take hours to read by hand.
- One workspace that covers research, drafting, and contract analysis together.
- Enterprise-grade data controls suited to confidential client matters.
Limits
Like all legal AI, Harvey can misstate authority or produce a draft that needs correction. It supports attorney judgment rather than replacing it, so a lawyer must review each output. Coverage also depends on the jurisdictions and sources connected to a given deployment, so results can vary by practice area and region. These are limits of the category, not flaws unique to Harvey, but firms should plan review steps into any workflow.
Harvey pricing
Harvey uses custom pricing. There are no public rates, and firms work with the Harvey sales team to scope a deployment based on size, use, and required features. This model is common for enterprise legal software, where seat counts and data needs vary across firms.
Because pricing is custom, the cost depends on how many attorneys use Harvey and which capabilities the firm turns on. Larger deployments with firm-wide onboarding and dedicated support sit at the higher end. Firms evaluating Harvey should ask for a scoped quote and a pilot so they can measure time saved against the seat cost before a full rollout.
Who should use Harvey?
Harvey fits firms and legal teams that do heavy research and drafting and want a secure, citation-grounded AI to speed that work. The clearest use cases are:
- Large law firms that need to research and draft across many matters at once.
- Litigation teams reviewing long filings, records, and case law.
- Transactional and corporate groups analyzing and comparing contracts.
- In-house legal departments that want a firm-grade assistant with strong data controls.
- Professional services teams handling regulatory and compliance research.
Harvey is less suited to solo practitioners or small firms with light research needs. The custom pricing and enterprise focus mean the tool delivers the most value where volume is high and the security and integration requirements are strict.
Alternatives and how it compares
Harvey competes with a growing set of legal AI tools. The main alternatives take different angles on research and drafting.
CoCounsel appeals to firms that want a legal AI backed by a deep research library. Spellbook targets transactional lawyers who want drafting help where they write. Harvey stands out for firms that want one workspace covering research, drafting, and analysis with enterprise data controls and citation grounding across the board. The right choice depends on whether a firm leans toward research depth, contract drafting, or a broad all-in-one assistant.
Limitations and getting started
Harvey has honest drawbacks worth weighing. Custom pricing means no upfront numbers and a sales conversation before you can start. The product aims at large firms, so smaller practices may pay for scale they do not use. And as with any legal AI, output needs attorney review, since the model can misstate a holding or cite authority that does not fit.
Getting started
- Contact the Harvey team to scope a deployment for your firm size and practice areas.
- Run a pilot with a small group of attorneys on live matters to measure time saved.
- Set review steps so every citation and draft gets checked by a lawyer before use.
- Configure access and data controls to match your confidentiality requirements.
- Expand to more attorneys once the workflow and value are proven.
A staged rollout keeps risk low and gives the firm evidence before a full commitment. With review built in and the right practice areas connected, Harvey can cut research and drafting time while keeping attorneys in control of the work.
Pros & cons
What we like
- Purpose-built for legal work with research, drafting, and contract analysis in one workspace
- Trusted by large law firms and professional services teams for demanding matters
- Handles long documents and multi-step legal questions with grounded citations
- Enterprise controls for data handling, access, and matter confidentiality
What could be better
- Custom pricing means no public rates and a sales process before you start
- Aimed at large firms, so smaller practices may find it more than they need
- Output requires attorney review, since legal AI can misstate authority
The verdict
Harvey is a strong legal AI assistant for large firms that want research, drafting, and contract analysis in one secure workspace, though its custom pricing and firm-scale focus make it a better fit for bigger practices than solo attorneys.