Keep pulling the thread on Gabe Pereyra.
Harvey is conceptualizing legal client matters, such as fund formations or litigations, as reinforcement learning (RL) environments for its AI agents to operate within.
To significantly improve AI models for complex legal work, access to the underlying decision-making process and "reasoning traces" is required, not just the final public documents like SEC filings.
Harvey's vision is to create a collaborative AI platform, analogous to Figma for design, that can be applied to law, code, and other professional verticals.
Harvey has signed Walmart and AT&T as customers, expanding its client base to include Fortune 500 companies and large private equity firms.
Law firms are concerned that AI tools like Harvey could fundamentally alter their hiring models, potentially reducing the number of associates needed to produce future partners.
Law firms are beginning to act as implementation partners for Harvey, offering to build workflows and manage deployment for their own in-house legal clients.
Harvey decided against building its own law firm after researching the failure of Atrium and concluding that simultaneously building a law firm and a tech company is too difficult.
Harvey's product strategy has shifted from focusing on individual lawyer productivity to enhancing the productivity and profitability of entire legal teams and law firms.
Harvey's long-term vision is to build a collaboration platform for the entire professional services industry, expanding beyond its initial focus on legal services.
Harvey is finding significant demand from large enterprise customers, such as major banks, for foundational services like building document management systems for their legal departments.
Harvey has established a deployed engineering force, similar to Sierra's agent engineering program, to work directly with customers on implementation and customization.
Gordon Moody, a former partner at the law firm Wachtel, serves as an advisor to Harvey.