Harvey AI: What It Does, Who Uses It and How It Compares

Last updated: 17 October 2026. Company figures are taken from Harvey’s own announcements and named press reports on the dates given. This guide is for information only and is not legal or purchasing advice.

Short answer: Harvey AI is a generative AI platform built for lawyers. Law firms, in-house teams and professional services firms use it to research, draft, review documents in bulk and run multi-step legal tasks through AI agents. Founded in 2022 in San Francisco, it raised $550 million at a $15.5 billion valuation in September 2026 and is the most heavily funded company in legal AI.

Key takeaways

• Harvey was founded in 2022 by Winston Weinberg, a former litigator at O’Melveny & Myers, and Gabe Pereyra, a former AI researcher at Google DeepMind and Meta. The OpenAI Startup Fund was an early backer.
• On 9 September 2026 Harvey announced a $550 million round at a $15.5 billion valuation, co-led by Diffusion and Lightspeed. Six months earlier, in March 2026, it was valued at $11 billion.
• Its core products include Assistant, Vault (bulk document analysis), Workflows and Agents, Knowledge (research) and add-ins for Word and Outlook.
• Main competitors are Legora, Thomson Reuters’ CoCounsel and LexisNexis’s Lexis+ AI with Protégé, although Harvey also partners with LexisNexis for content.
• Harvey does not publish its prices. It sells enterprise contracts, so cost depends on seats, products and term.

What is Harvey AI?

Harvey is software, not a law firm and not a chatbot for consumers. It sits on top of large language models and adds the things a legal team needs before it can trust AI with client work: secure document storage, citations back to the source text, legal research content, permissions, and tools designed around how lawyers actually work, such as due diligence tables and drafting in Word.

The company began when its two founders tested an early prompt on California tenant law and found that lawyers approved most of the answers. That proof of concept, plus early access to OpenAI’s models, helped Harvey win its first large firm. In February 2023 Allen & Overy (now A&O Shearman) announced it was rolling Harvey out across the firm after a trial involving about 3,500 of its lawyers. That deal made Harvey the name most associated with “AI for Big Law”.

Since then Harvey has moved from a single assistant to a broad platform. It now describes its aim as helping legal teams “own their intelligence”, meaning firms should build their own know-how, agents and even models on top of Harvey rather than relying on generic AI tools. If you are new to the wider category, our overview of legal AI software and assistants sets out where Harvey fits.

What does Harvey AI do? The main products

Harvey’s product line changes quickly, so treat this as a snapshot as of October 2026, based on the company’s own product pages.

Harvey also lists newer modules such as Spaces (shared workspaces with clients or co-counsel), Contract Intelligence, Horizon Scanning for regulatory change, Intake for in-house request management and Command Center for adoption analytics. In March 2026 the company said customers were running more than 25,000 custom agents on the platform.

A notable change in 2026 was Harvey’s first in-house model, Harvey Tenet. According to TechCrunch, it was built by post-training an open-weight model on legal data, and Harvey now encourages customers to post-train open-weight models of their own rather than depend entirely on the big proprietary AI labs.

Who uses Harvey AI?

Harvey’s customers are mostly large organisations. In its September 2026 funding announcement, Harvey said 80% of Am Law 100 firms use it, along with in-house legal teams at five of the Fortune 10. Firms it has named publicly include A&O Shearman, Ashurst, Baker Donelson, Cuatrecasas and GSK Stockmann, and PwC adopted it for its legal business.

The common thread is volume and value. Harvey is built for teams that review thousands of documents, handle high-stakes transactions or litigation, and can afford enterprise software with a dedicated rollout. Solo practitioners and small firms are not its core market, and most will find cheaper tools that fit their needs. Before signing anything, read our guides on what legal AI actually costs and how to run a legal AI pilot that produces a decision.

How much has Harvey raised, and what is it worth?

Harvey has raised money at a pace unusual even for AI start-ups. TechCrunch reported that the September 2026 round took total funding above $1.55 billion and that the valuation nearly doubled in about nine months, from the $8 billion it reported for a December 2025 round.

Sources: Harvey’s announcements of the $11 billion round and the $15.5 billion round, and TechCrunch’s report. Other backers include Kleiner Perkins, Andreessen Horowitz, Coatue, Conviction, Elad Gil and Goldman Sachs Alternatives.

Harvey does not publish full revenue figures. When it announced its acquisition of Benchmark in July 2026, it said it had added more than $100 million in net-new annual recurring revenue in the second quarter alone. It has also been buying companies: Lume and Hexus in March 2026, Benchmark in July and Guardrails AI, an AI agent security company, in September, which it called its fourth acquisition of the year. We covered the Guardrails deal and what these valuations mean for buyers in legal AI funding in 2026.

How does Harvey compare with CoCounsel, Lexis+ AI and Legora?

There is no single “best” legal AI tool. The right choice depends on whether you need research content, document review at scale, collaboration or deep integration with existing systems.

A few facts help frame the market. Thomson Reuters built CoCounsel on its 2023 acquisition of Casetext and has since expanded CoCounsel Legal, including a UK launch in January 2026. Legora raised a $550 million Series D at a $5.55 billion valuation in March 2026, led by Accel, and reports more than 800 customers in over 50 markets. And in June 2025 LexisNexis and Harvey announced a strategic alliance that brings Lexis primary law content and Shepard’s Citations into Harvey. So LexisNexis is both a competitor and a partner.

For a closer look at how document review tools differ, see our guide to AI contract review software.

What are the risks of using Harvey or any legal AI?

Harvey’s own customers stress that output must be checked. The professional risk sits with the lawyer, not the vendor. Courts have already sanctioned lawyers who filed AI-generated citations to cases that did not exist, most famously in Mata v. Avianca in New York in 2023.

• Accuracy: citations and quotations need to be checked against the source, every time.
• Confidentiality: confirm where data is stored, whether it is used to train models, and who can access each matter.
• Ethics: in the US, ABA Formal Opinion 512 sets out duties of competence, confidentiality, supervision and reasonable fees when lawyers use generative AI.
• Billing: if a task takes a fraction of the time, clients will expect bills to reflect that.
• Lock-in: check what happens to your vaults, agents and data if you leave.

Is Harvey AI worth it?

For a large firm or legal department with heavy document work, a serious budget and a team to drive adoption, Harvey is one of the leading options, and its funding means it is unlikely to disappear soon. For smaller practices the case is weaker: the price is set for enterprise buyers and much of the value comes from scale. Either way, run a structured pilot on your own matters and compare at least two tools before committing. The work Harvey automates is also changing junior roles; we look at that in should junior lawyers become legal engineers?

Frequently asked questions

Who owns Harvey AI?

Harvey is a private company. It is led by co-founders Winston Weinberg (CEO) and Gabe Pereyra, and backed by venture and growth investors including Sequoia, Kleiner Perkins, Andreessen Horowitz, GIC, Lightspeed and Diffusion. It is not owned by OpenAI, although the OpenAI Startup Fund was an early investor.

How much does Harvey AI cost?

Harvey does not publish prices. It sells annual enterprise contracts, and cost depends on the number of users, the products included and the length of the deal. Ask for a written quote and a pilot.

Is Harvey AI a law firm?

No. Harvey is a software company. It does not give legal advice to the public, and lawyers using it remain responsible for the work they deliver.

What AI model does Harvey use?

Harvey has historically built on models from leading AI labs, including OpenAI, and adds its own legal tuning and tools. In 2026 it released Harvey Tenet, its first in-house model, built by post-training an open-weight model on legal data.

Can individuals or small firms sign up for Harvey?

Harvey is aimed at large firms, in-house teams and professional services organisations. Small firms can contact its sales team, but many will find better value in tools designed for smaller practices.

What is Harvey Vault?

Vault is Harvey’s secure document workspace. Each vault holds up to 100,000 files and lets users ask questions across all of them or build review tables that pull the same information from every document.

Legal Desire has no commercial relationship with Harvey or the other vendors named. Funding figures and product details change quickly; check the company’s own announcements before relying on them.