Harvey
by Harvey
Generative AI platform for law firms, in-house legal teams and professional services
Harvey is an enterprise generative AI platform for law firms, corporate legal departments and professional-services firms. It handles legal research, drafting, contract review and bulk document analysis, and runs multi-step legal agents, so lawyers spend less time on first drafts, diligence and review while keeping client data under enterprise security controls.
Harvey is a generative AI platform built for legal and professional-services work, founded in 2022 by Winston Weinberg, a former securities and antitrust litigator at O'Melveny & Myers, and Gabriel Pereyra, a former research scientist at Google DeepMind and Meta, and headquartered in San Francisco. It covers legal research, drafting and document analysis through several surfaces: Vault for securely storing and bulk-analysing large document sets, Knowledge for legal, regulatory and tax research, Spaces for matter-based collaboration across organisations, Contract Intelligence for review and negotiation, purpose-built Agents that run multi-step legal tasks end to end, and Command Center analytics for measuring adoption. Harvey II, launched on 18 August 2026, put Memory at the centre, carrying each lawyer's preferences across Harvey, Word, Outlook and agents; ethical-wall enforcement arrived through an Intapp partnership in July, and an Everlaw eDiscovery integration is due for joint customers in fall 2026. Harvey largely rents its underlying intelligence from model makers such as OpenAI and Anthropic, contractually requiring zero data retention from them, and in August 2026 previewed Harvey Tenet, an open-weight Kimi K3 base post-trained with Fireworks for long-horizon legal work. In May 2026 it released Legal Agent Bench, an open-source benchmark of more than 1,200 tasks across 24 practice areas. Harvey says more than 2,400 legal organisations and 200,000 professionals in over 70 countries use it, including 80% of the Am Law 100 and in-house teams at five of the Fortune 10, and Microsoft's legal group (CELA) adopted it in July 2026. On 9 September 2026 it raised $550 million at a $15.5 billion valuation, co-led by Diffusion and Lightspeed, six months after a $200 million round at $11 billion. Independent coverage names Legora and Thomson Reuters as its main competitors.
The chief innovation or knowledge-management officer at a large law firm, or the general counsel of a large enterprise legal department, who needs one secure AI platform that lawyers across practice groups will actually adopt.
Faster first-pass research, drafting, diligence and contract review on a platform Harvey says is used by 80% of the Am Law 100, with firm-grade confidentiality controls.
At a Glance
- Category
- Industry & Government
- Pricing
- Contact for pricing
- Target Market
- General Counsel, Law Firm Leaders, Chief Innovation Officers, CIOs
- Deployment
- Cloud-first
- Founded
- 2022
- Headquarters
- San Francisco, United States
- Customers
- 2,400+ legal organizations and 200,000+ professionals in 70+ countries (per Harvey, September 2026)
Key Features
- ✓Vault
Securely stores, organises and bulk-analyses large legal document sets, with agentic search added in September 2026, so diligence and review scale beyond what a team can read manually.
- ✓Knowledge
Researches complex legal, regulatory and tax questions across domains, giving lawyers a sourced starting point for memos and advice instead of a blank page.
- ✓Agents
Purpose-built agents execute multi-step legal work end to end, including contract review agents launched in September 2026, reducing hand-offs on routine matters.
- ✓Memory (Harvey II)
Learns and retains how each lawyer works and carries those preferences across Harvey, Word, Outlook and agents, so outputs come back consistent without repeated prompting.
- ✓Spaces
Matter-based shared workspaces let legal teams collaborate securely across organisations, keeping documents, tasks, permissions and work history together in one place.
- ✓Contract Intelligence
Surfaces insights across contracts to strengthen negotiations and accelerate reviews, useful wherever contract volume outpaces the capacity of the reviewing team.
- ✓Command Center
Analytics and benchmarking that measure AI adoption across the organisation, helping leaders see where the platform is actually used and justify the spend.
- ✓Enterprise security controls
SAML SSO, audit logs, IP allow-listing, regional data residency and a contractual no-training guarantee address the confidentiality questions firms raise first.
Capabilities
Use Cases
- •M&A due diligence
Deal teams load data-room documents into Vault and use bulk analysis and agents to flag key provisions and issues faster than a manual page-by-page review.
- •Litigation and eDiscovery preparation
Litigators build chronologies and prepare depositions from evidence, with the Everlaw integration (expected fall 2026) querying millions of documents and citing back to source.
- •Contract review and negotiation
In-house teams run contract review agents and Contract Intelligence to speed first-pass review and strengthen negotiating positions across large volumes of agreements.
- •Regulatory and tax research
Advisers research complex legal, regulatory and tax questions in Knowledge and produce sourced first drafts of memos that a qualified lawyer then verifies before use.
- •Firmwide AI rollout and governance
Knowledge-management and innovation leaders deploy Harvey across practice groups, enforce ethical walls through Intapp, and use Command Center to track adoption ahead of renewal.
Ideal For
Best For
- ✓Am Law 100 and large international law firms rolling out generative AI firmwide
- ✓Corporate legal departments at large enterprises handling high contract volumes
- ✓Transactional teams running M&A due diligence over large document sets
- ✓Litigation, regulatory and tax practices that need sourced research and first drafts
- ✓Professional-services firms that require strict confidentiality, regional data residency and ethical-wall controls
Not Ideal For
- ✗Solo practitioners and small firms that want transparent, upfront pricing: Harvey publishes no price list and access runs through a sales demo, and Lawyerist lists cost as a con.
- ✗Teams that only need a basic, low-cost drafting assistant, which Lawyerist identifies as a poor fit for Harvey.
- ✗Firms without capacity to invest in implementation, training and lawyer review of every output, since the lawyer remains responsible for verifying what the system produces.
Deployment
Market & Ratings
2,400+ legal organizations and 200,000+ professionals in 70+ countries (per Harvey, September 2026)
Market Analysis
Pros
- ✓Legal-specific platform covering research, drafting, document analysis and agents in one product (Lawyerist)
- ✓Deep adoption at the top of the market: Harvey reports use at 80% of the Am Law 100 and by in-house teams at five of the Fortune 10, and Microsoft's legal group adopted it in July 2026
- ✓Strong security posture: SOC 2 Type II, ISO 27001, ISO 27701 and ISO 42001, regional hosting, zero data retention required of model providers, and a contractual no-training guarantee
- ✓Fast growth and heavy funding: The Next Web reported annualised revenue above $350M in August 2026, up from $190M in January, and Harvey raised $550M in September 2026
- ✓Works inside existing tools, with Word, Outlook and iManage integrations (Lawyerist) plus Intapp and Everlaw partnerships
Cons
- ✗No published pricing; firms must request a demo or contact sales, and Lawyerist lists the product as expensive
- ✗Lawyerist classes it as an add-on product, so it typically sits on top of existing research and document systems rather than replacing them
- ✗Dependence on third-party model makers such as OpenAI and Anthropic; The Next Web argued a roughly 44x revenue multiple leaves little room if model owners squeeze app builders
- ✗Hacker News practitioners have questioned client confidentiality when documents are processed by an outside model provider, and the defensibility of a product built on GPT-4
- ✗Accuracy evidence is largely vendor-produced: Legal Agent Bench was built by Harvey and covered without independent scrutiny, and the widely cited Stanford legal-AI hallucination study tested Lexis+ AI and Westlaw rather than Harvey
Pricing
Enterprise
Contact for pricing
- ✓Vault, Knowledge, Agents, Spaces and Contract Intelligence
- ✓SAML SSO, audit logs and IP allow-listing
- ✓Regional data residency (EU or Switzerland, US, or Australia)
- ✓Contractual guarantee that customer data is not used to train models
Harvey does not publish list pricing: harvey.ai has no pricing page, and Lawyerist notes firms must request a demo or contact Harvey for a quote, listing cost as a con. Per-seat figures circulating on competitor and SEO blogs cite no primary source and are not vendor-confirmed, so none are reproduced here; expect a negotiated enterprise contract and ask sales about seat minimums, term length and any add-on content fees.
Security & Compliance
Sources
This page was written from 15 sources, 8 on domains other than harvey.ai.
- 1.harvey.ai — harvey.aivendor
- 2.harvey.ai — blogvendor
- 3.harvey.ai — harvey raises dollar550m at a dollar155b valuation to help lvendor
- 4.harvey.ai — harvey raises at dollar11 billion valuation to scale agents vendor
- 5.harvey.ai — securityvendor
- 6.harvey.ai — post training update harvey tenetvendor
- 7.harvey.ai — harvey everlaw evidence ediscoveryvendor
- 8.thenextweb.com — harvey legal ai 15 5bn valuation 500m raise vertical ai
- 9.artificiallawyer.com — next gen harvey ii launches with memory at its core
- 10.artificiallawyer.com — harvey deepens microsoft relationship
- 11.artificiallawyer.com — harvey launches legal agent bench
- 12.lawyerist.com — harvey ai review artificial intelligence for lawyers
- 13.en.wikipedia.org — Harvey (software)
- 14.hn.algolia.com — 35735030
- 15.hai.stanford.edu — ai trial legal models hallucinate 1 out 6 or more benchmarki
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
Faro Health
Agentic AI that turns clinical trial protocols from static documents into structured development intent
Lumafield Quality Agent
Physical-AI quality agent that catches unanticipated manufacturing defects and traces root causes from industrial CT and sensor data
Sixfold
An AI underwriter that learns your book and takes submissions through to bind-ready
Casepoint IQ
Governed agentic AI for eDiscovery, FOIA and investigations, on FedRAMP High and DoD IL6 infrastructure