Harveya startup developing an AI-powered “copilot” for lawyers has raised $100 million in a Series C funding round led by GV, Google’s corporate enterprise arm.
With the tranche, which also included participation from influential angel investors and enterprise capitalists OpenAI, Kleiner Perkins, Sequoia Capital, Elad Gil and SV Angel, Harvey has raised a complete of $206 million and the corporate's value is $1.5 billion.
In a post On Tuesday, co-founders Winston Weinberg and Gabriel Pereyra explained on Harvey's official blog that the vast majority of the brand new capital will go toward collecting and curating data to construct and train “domain-specific” AI models, while also increasing Harvey's headcount and expanding its paid services to recent geographic areas.
“This investment will enable Harvey to proceed to scale and enhance our AI-powered technology across all business functions and geographies,” Weinberg and Pereyra said. “We will use this recent capital to speculate within the engineering, data and domain expertise which are fundamental to constructing AI-native systems that enable probably the most complex knowledge work. We can even deepen our partnerships with cloud and model providers to integrate additional models into Harvey and expand our training collaborations to further improve model effectiveness.”
Weinberg, a former securities and antitrust litigator on the law firm O'Melveny & Myers, and Pereyra, previously a scientist at DeepMind, Google Brain (one other Google AI group) and Meta AI, founded San Francisco-based Harvey in 2022. Weinberg and Pereyra are roommates; Pereyra showed Weinberg OpenAI's GPT-3 text generation system and Weinberg realized it could possibly be used to enhance legal workflows.
Harvey, based on OpenAI's GPT-4 family of models, can answer legal questions phrased in natural language, equivalent to “Tell me what the differences are between an worker and an independent contractor within the Fourth Judicial Circuit” and “Tell me if this clause in a lease violates California law, and if that’s the case, rewrite it in order that it not violates it.” Harvey also offers tools that robotically extract information from court transcripts, robotically find legal documents that may support court arguments, and create initial drafts of briefs that include information and citations from legal databases.
In theory, that is powerful stuff. But additionally it is tricky. Given the sensitive nature of most litigation, some lawyers and law firms is perhaps hesitant to provide a tool like Harvey access to case documents. There can be the problem of the propensity of language models to toxicity And invented factswhich can be particularly badly received in court – if not even considered perjury.
For this reason, Harvey comes with a disclaimer: The tool will not be intended as legal advice for non-lawyers and may only be used under the supervision of licensed attorneys.
Harvey has competition. Casetext uses AI, primarily OpenAI models, to seek out legal cases and help with general legal research tasks and transient drafting. More surgical tools like Klarity use AI to simplify the tedious task of contract review. Startup Augrented even at one point explored ways to make use of OpenAI models to summarize legal notices or other sources in plain language to assist apartment tenants defend their rights.
However, Weinberg and Pereyra claim that Harvey is on a roll and is used “on daily basis” by tens of 1000’s of lawyers at law firms and consulting firms equivalent to Allen & Overy, Macfarlanes, Ashurst, CMS, Reed Smith and PwC. Annual recurring revenue has tripled since last December, the 2 co-founders said within the blog post, while Harvey's workforce has tripled.
The information reported In early June, Harvey had hoped to boost $600 million at a valuation of “at the least” $2 billion, partially to accumulate a legal research service called vLex to coach its AI products. These plans fall through – hence the greatly reduced Series C.
