Harvey, a startup developing an AI-based “pilot” for lawyers, has raised $100 million in a Series C round led by GV, Google’s corporate venture arm.
This round of funding also saw participation from large angel investors and VCs including OpenAI, Kleiner Perkins, Sequoia Capital, Elad Gill, and SV Angel, bringing Harvey's total raised to $206 million, valuing the company at $1.5 billion.
In a post on Harvey's official blog on Tuesday, co-founders Winston Weinberg and Gabriel Pereyra said much of the new capital will be used to grow Harvey's headcount, expand its paid services into new geographies, and focus on collecting and curating data to build and train “domain-specific” AI models.
“This investment will enable Harvey to scale and improve its AI-based technology across business functions and geographies,” Weinberg and Pereyra said. “We will use this new capital to invest in the engineering, data, and domain expertise that are fundamental to building AI-based systems that facilitate the most complex knowledge work. We will also deepen our partnerships with cloud and model providers to integrate additional models into Harvey and expand training collaborations to continually improve model efficacy.”
Weinberg, a former securities and antitrust litigator at law firm O'Melveny & Myers, and Pereyra, a former research scientist at DeepMind, Google Brain (another Google AI group), and Meta AI, founded Harvey in San Francisco in 2022. Weinberg and Pereyra are roommates. Pereyra showed Weinberg OpenAI's GPT-3 text-generation system, and Weinberg realized that it could be used to improve legal workflows.
Powered by OpenAI’s GPT-4 family of models, Harvey can answer legal questions expressed in natural language, such as “What is the difference between an employee and an independent contractor in the Fourth Circuit?” and “Tell me if this clause in my lease violates California law, and if so, rewrite it so that it no longer violates it.” Harvey also provides tools to automatically extract information from court records, automatically find legal documents that support court arguments, and generate drafts of submissions that incorporate information and citations from legal databases.
In theory, it’s powerful. But it’s also risky. Given the sensitive nature of most legal disputes, some lawyers and law firms may be reluctant to give tools like Harvey access to all their case documents. Language models also tend to churn out toxicity and falsehoods, which is especially bad in court—even if it’s not perjury.
That's why Harvey has the following disclaimer: This tool is not intended to provide legal advice to non-attorneys and should only be used under the supervision of a licensed attorney.
Harvey faces competition. Casetext primarily uses OpenAI models to find legal cases and assist with general legal research tasks and brief drafts. Surgical tools like Klarity use AI to remove tedious tasks from contract review. At one point, startup Augrented even explored ways to use OpenAI models to summarize legal notices or other sources in plain language to help apartment tenants defend their rights.
But Weinberg and Pereira claim Harvey is flying high, used “every day” by tens of thousands of lawyers at law and consulting firms including Allen & Overy, MacFarlane, Ashurst, CMS, Reed Smith and PwC. Annual recurring revenue has tripled since December, the two co-founders said in a blog post, and Harvey’s staff has also tripled in size.
The Information reported in early June that Harvey had hoped to raise $600 million at a valuation of “at least” $2 billion, some of which would go toward acquiring a legal research service called vLex to train its AI products. That plan fell through, and so the Series C was significantly scaled back.
