Why OpenAI and Google Are Betting on This AI Unicorn with a $100 Million Deal

On July 4, 2022, Winston Weinberg called the entire OpenAI executive team. Two weeks earlier, the 27-year-old and his friend Gabe Pereyra had emailed the company about a chatbot they had developed that answered legal questions from a public online forum. The tool, with an 86% success rate, showed enough promise that OpenAI executives could work through the holidays.

The July 4th call kicked off Harvey’s AI legal platform, which lets lawyers upload files and enter requests to automate tasks like document analysis, consolidating lengthy research, and even tracking productivity. Last month, the company closed a $100 million Series C funding round with investments from Google Ventures, OpenAI, Kleiner Perkins, Sequoia Capital, and tech investor Elad Gil, bringing its total funding to $216 million. The new deal values ​​Harvey at $1.5 billion—the highest-valued startup in OpenAI’s portfolio.

“I had been working as a lawyer for a little less than a year and realized that many of the jobs my colleagues were doing were jobs they could have done before they graduated from law school,” says Weinberg, a member of the Forbes Under 30 list.

After graduating from law school, Weinberg began using ChatGPT for mundane, routine tasks assigned to new associates, such as picking apart thousands of pages of documents. That’s when he approached former Google AI scientist Pereyra—whom he met through mutual friends and became roommates with—with an idea: automate legal tasks so associates like him could focus on more meaningful work, like crafting thoughtful arguments for their clients.

Weinberg says he was working more than 120 hours a week — moonlighting on an artificial intelligence idea while working full-time as a litigation lawyer — before he publicly declared Hurricane Harvey was coming in late 2022.

After launching, Harvey announced that its first client was the billion-dollar law firm A&O Shearman. The second was the Big Four accounting firm PwC. Unlike many business-oriented firms that start by choosing smaller companies as their first clients, Harvey focused on the biggest players in the game.

“One thing I was very confident about that a lot of people didn’t believe me about was to go after the largest, most prestigious companies first,” says Weinberg, a 2024 Forbes Under 30 graduate. “The biggest thing with AI is going to be trust, and gaining the trust of these larger institutions early on is key, given that we’re helping these people do incredibly high-profile work.”

To land such clients, Weinberg says he found the most recent public legal document filed by a potential client and generated potential counterarguments that could be used against them in court, using Harvey’s model. The personalized approach worked—and all it took was the trust of one or two big names to get things going. Since then, they’ve hired law firms like O’Melveny & Myers, Vinson & Elkins, Gleiss Lutz, ReedSmith, Macfarlanes and others.

That’s a strong start. But AI is still relatively new and unregulated, potentially raising issues like violating confidentiality or generating false information. Recent headlines — like Donald Trump’s former lawyer Michael Cohen filing a lawsuit with bogus AI-generated court orders or rapper Pras seeking a retrial because his previous lawyer misused the tool — are forcing law firms and startups to be more cautious about adopting the technology, says legal tech strategist Nicole Black.

To overcome the obstacles of AI in the legal profession, Weinberg has turned to industry veterans like longtime Google lawyers Andrew Hyman and John Labarre, who have joined Harvey as general counsel, and former Wachtell partner Gordon Moodie as Harvey’s chief product officer. Weinberg says the latest round of funding will help bring in more high-quality lawyers and engineers to continue training their model to be more specialized for law firms while maintaining accuracy and data privacy.

Still, there is competition in the legal AI space. While Harvey’s offering has advantages over narrower legal tech startups like Spellbook—which focuses solely on contracts—direct competitors like Casetext, which Thomson Reuters bought for $650 million last year, pose a threat. But among the thousands of startups developing software using Open AI technology, Open AI COO Brad Lightcap says the speed and scale of Harvey’s vision stand out.

“You give them some context and advice, and a week later they come back 20 times smarter,” Lightcap says. “They don’t see themselves as people selling software, they see themselves as partners in the industry.”

For Sarah Guo, one of Harvey’s early investors and a member of the Forbes Under 30 list, the co-founders’ optimism about the value of AI far outweighs any challenges.

“(Imagine) taking on more client work because you’ve automated so much of your juniors’ work, or winning cases more often because the discovery process takes a day instead of months,” Guo says. “At some point, AI will impact your productivity so much and give you a competitive advantage that you won’t be able to avoid it.”