Legal AI: The Rise of 'Harvey'

Introduction to Legal AI and Harvey

Legal AI is the application of artificial intelligence technologies to simulate human intellectual processes in the legal field. It aims to automate routine legal tasks, improve productivity in document review, legal research, and contract analysis, helping legal teams save time in finding legal information (Thomson Reuters, Bloomberg Law). According to U.S. law (15 U.S. Code § 9401), AI is defined as "a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments" (Cornell Law School). "Harvey," a startup at the forefront of Legal AI, is one of Silicon Valley's most promising companies and has successfully attracted major investors. Its list of backers includes prominent names like the OpenAI Startup Fund, Sequoia Capital, Kleiner Perkins, Elad Gil, Google Ventures, Coatue, and most recently, Andreessen Horowitz.

Harvey's Rapid Growth and Market Expansion

The value of "Harvey," headquartered in San Francisco, has seen a significant and remarkable increase, jumping from $3 billion in February 2025 to $5 billion in June, then reaching $8 billion in late October. This rapid growth reflects the high valuations private investment firms are giving to AI companies, in addition to Harvey's ability to earn the trust of major law firms and in-house legal departments. The startup now claims to serve 235 clients in 63 countries, including the majority of the top 10 U.S. law firms. Its annual recurring revenue also exceeded $100 million as of August.

The Genesis and Development of Harvey's Idea

The idea for "Harvey" began when co-founder Winston Weinberg, working as a legal assistant at "O'Melveny & Myers," discovered the transformative potential of GPT-3 in legal work. Collaborating with co-founder Gabe Pereira, who was working at Facebook at the time, they developed a "chain-of-thought" approach using GPT-3 to answer legal inquiries. Their experiment showed that 86% of the AI-generated answers were acceptable to real lawyers without modification. Subsequently, the co-founders contacted Sam Altman and Jason Kwon of OpenAI via email. As a result, the OpenAI Startup Fund invested in "Harvey" and introduced it to prominent angel investors such as Sarah Guo and Elad Gil.

Operational Strategy and Challenges

Weinberg emphasizes the importance of strong company performance to attract investment and believes that the best way to raise funding is to ensure excellent operational execution. The company's annual recurring revenue reached $100 million in August, with a significant portion of the costs attributed to high computing expenses, especially as the company operates in over 60 countries and adheres to data residency laws in each.

via GIPHY

Revenue Model and Competitive Advantage

In early 2025, 4% of "Harvey's" revenue came from corporations and 96% from law firms. Now, corporate revenue has risen to 33% and is expected to reach 40% by year-end. Harvey's sales strategy relies on demonstrating how its product can support law firms in their work, and law firms themselves often help persuade other companies to collaborate with Harvey. Harvey is moving towards developing a "multi-player" platform that allows in-house lawyers to securely communicate with external parties, considering ethical walls and data permissions across dozens of countries. This is a key competitive advantage, as there is currently no competitor offering a comprehensive platform that connects legal service providers and consumers.

Harvey's Uses and the Future of Legal AI

Lawyers primarily use "Harvey" in areas such as: legal drafting, research (thanks to a strategic partnership with LexisNexis), and analysis (such as conducting 10 questions on 100,000 documents in due diligence or legal discovery processes). Initially, use cases focused on transactions like mergers and acquisitions and fund formation, but the litigation field is experiencing rapid growth. Weinberg emphasizes that "Harvey" is not just a ChatGPT interface. He believes the company's biggest competitive advantage lies in collecting vast amounts of legal workflow data, developing robust assessment frameworks for AI output quality, and building a true "multi-player" platform that connects legal service providers and consumers. Harvey's business model currently relies on per-user fees, but it is moving towards results-based pricing as workflow complexity increases. Weinberg sees AI as a powerful educational and training tool for junior lawyers, helping them become partners as quickly as possible. Harvey does not plan to raise large funding rounds in the near future, as it does not need a lot of money and does not burn through much. However, the company is interested in public markets in the long term.