Thomson Reuters agreed to buy Casetext for $650 million in cash, placing a legal AI startup at the center of its plans to expand generative AI offerings. The acquisition was expected to close in the second half of 2023, subject to regulatory approvals and customary closing conditions.
From legal community to AI tools
Casetext was founded in 2013. It first focused on building a community where attorneys could share knowledge, alongside a service that gave users free access to legal texts annotated by lawyers.
The company later shifted toward artificial intelligence and machine learning, developing automated workflows and tools for legal teams. Its flagship product, CoCounsel, uses AI to review documents, assist with legal research memos, prepare depositions and analyze contracts.
Casetext had early access to OpenAI’s GPT-4 language model, which provides the infrastructure behind CoCounsel. That connection positioned the company to offer generative AI capabilities tailored to legal work.
A deal with a broad customer base
At the time of the acquisition announcement, Casetext had 104 employees and more than 10,000 law firms and corporate legal departments as customers, according to a press release. The startup had raised more than $64 million from Union Square Ventures and other investors.
Thomson Reuters CEO Steve Hasker described the purchase as part of a “build, partner and buy” strategy for bringing generative AI to customers. He said the acquisition could accelerate and expand the company’s market potential for those offerings and change how professionals work.
Casetext CEO Jake Heller said the company had spent ten years using AI to build products intended to elevate legal practice and help attorneys serve more people’s legal needs. He described joining Thomson Reuters as an opportunity to advance that mission and generative AI solutions across professions.
Thomson Reuters’ wider AI plans
The acquisition fits a larger plan to embed generative AI in Thomson Reuters’ legal, tax, accounting and news businesses. The company had announced plans to spend about $100 million a year on AI and incorporate generative AI into products in the second half of that year.
It also planned to set aside $10 billion for mergers and acquisitions from then until 2025, with many of those deals focused on AI. Buying Casetext gave Thomson Reuters an established legal technology business and a product already built around AI-assisted workflows.
The transaction therefore connected two parts of the company’s strategy: developing AI capabilities and acquiring businesses that could bring those capabilities into professional tools. The stated ambition extended beyond lawyers to customers across Thomson Reuters’ other business areas.
Interest and uncertainty around legal AI
A Reuters poll published in April 2023 found that 82% of legal professionals believed generative AI could readily be applied to legal work. A smaller majority, 51%, said it should be applied there. The article cautioned that the poll should be viewed with some skepticism because Reuters was reporting on its own survey.
Those figures suggest interest in potential uses, while the difference between the two responses shows that seeing an application and supporting its use are not the same thing. CoCounsel’s tasks—such as reviewing documents and analyzing contracts—illustrate the kinds of legal work where companies are attempting to apply the technology.
The broader outlook remained contested. Some observers saw the potential for generative AI to add trillions in value to the global economy; others raised concerns about hype and possible job losses. The article also noted a forecast that the generative AI market could generate $36 billion in revenue by 2028.
Against that backdrop, the Casetext purchase represented a significant commitment by Thomson Reuters to AI in legal services. Whether the technology delivers on its promise depends on how well these tools work in professional settings and how organizations choose to use them.