Simply sign up to the Technology sector myFT Digest -- delivered directly to your inbox.
Legal AI start-up Harvey has hit a valuation of $11bn and attracted half the biggest law firms in the US as customers. Fast-growing rival Legora has a celebrity spokesman, the actor Jude Law. But another AI business has dominated legal industry headlines in recent months.
Anthropic delivered a warning shot to legal tech specialists in late January, when it unveiled a tool that promises to automate routine work, such as contract reviews. The news rattled investors and sent share prices of companies including Thomson Reuters and LexisNexis owner Relx tumbling.
Just three months later, Anthropic extended its push into the sector, releasing a dozen new plug-ins and linking its Claude AI assistant to more than 20 software platforms used by the legal industry. Lawyers, tech experts and advisers sense that this is just the start.
"Everyone wakes up these days, and they've got one eye on Anthropic's blog and one eye on everything else," says Rudy DeFelice, global head of AI strategy at Harbor, a legal sector consulting firm based in Chicago. "That's kind of the frame through which they see the world."
Everyone wakes up these days, and they've got one eye on Anthropic's blog and one eye on everything else
As competition intensifies, incumbent legal tech companies are upgrading products and emphasising their selling points: troves of authoritative content and expertise that legal teams can trust. And rather than trying to fend off Anthropic, they are integrating its Claude tools into their platforms. Both Thomson Reuters and LexisNexis have expanded their partnerships with Anthropic.
"As the [AI] models get better and stronger, our products get better and stronger," says Sean Fitzpatrick, chief executive officer of LexisNexis's global legal business. But, as the technology advances at a rapid clip, "we need to run really fast to keep up".
The incumbents say the rise of AI poses an opportunity, not an existential threat. LexisNexis, which joined forces with Harvey as well in mid-2025, aims to give all lawyers "their own personal" AI assistants tailored to their specific needs, he says.
"It's going to understand their practice area, jurisdiction, preferences and style, and it's going to be able to do things for them in an automated way that they used to do manually," says Fitzpatrick. "What keeps me awake at night is turning that vision into a reality."
Thomson Reuters, which announced in February that its CoCounsel tool had attracted 1mn users, argued in Fortune this year that Claude is no substitute for platforms that can help research unresolved legal issues, validate agreements and develop novel arguments. The company has sought to distinguish itself from the new AI models with a focus on accuracy, accountability and reliability.
"One of the things that we're trying to counter is that folks will say, 'You know what? Using these other models, they're good enough'," says Laura Clayton McDonnell, president of the corporates business at Thomson Reuters in New York.
"Customers are going to have so many different choices," she adds. "And I think the choice that they'll ultimately lean to is 'fiduciary-grade AI', to help drive business, reduce risk, improve productivity and unlock growth."
Chief legal officers, at least in the short term, will still rely on specialist tools offered by the more established players because of the essential content and human judgment they provide, says Chris Audet, analyst at research firm Gartner.
Concerns over privacy and accuracy will also drive demand for those products. A federal judge in New York this year ruled a defendant's conversations with an AI chatbot that he later shared with his lawyers are not protected by attorney-client privilege. Separately, Sullivan & Cromwell apologised in April to a federal judge for a bankruptcy filing that contained several errors generated by AI. UK law firm Pinsent Masons was admonished in May by London's High Court after its lawyers made false submissions to a judge based on AI, misleading the court twice over a routine insolvency application.
But legal professionals -- who Anthropic says are the most engaged users of the Claude Cowork platform -- face increasingly tough questions over how to allocate limited budgets and whether new AI models will be sufficient, Audet says.
It is difficult to predict how long legal tech stalwarts will sustain their advantage given how rapidly the market is evolving.
Harvey and Legora are already shaking up the sector. Anthropic is shifting into another gear with new features and partnerships, including one with UK-based law firm Freshfields in April to build specialist AI tools including some that may be sold to rival law firms. Kirkland & Ellis, the world's highest-grossing law firm, has set aside about $500mn to create its own AI platform, on top of its spending on outside tools based on the technology.
Proprietary data and expertise developed over decades have given the legal tech leaders protection, providing a "moat" that is hard for challengers to breach, says Harbor's DeFelice. But these businesses, long seen as fortresses, are now facing doubts about their defences. "Over the next year or two, their moat is pretty solid . . . Over five years, I think it's harder to see."
Ultimately, the winners will probably combine the best elements of today's players, he adds. That means the "trust and content" of incumbents, the "speed and product imagination" of companies such as Harvey and Legora, and the "raw intelligence" of the AI models such as Claude.
Rather than depending on a single product, legal teams are moving towards a hybrid model, where they turn to different providers depending on the problems they face, he says.
There's not some huge market for streamlining the review of political advertising -- but it's huge for us
Cox Media Group is turning to "off-label AI", by working with providers to adapt tools for unexpected purposes. In one case, its lawyers can use a contract-processing tool to review political ads. These are a key source of revenue but also involve legal risks, says Eric D Greenberg, general counsel of Cox Media, which operates TV and radio stations in the US. "There's not some huge market for streamlining the review of political advertising -- but it's huge for us," he says.
Cox is also creating a "digital twin" of one of its most experienced lawyers. By mining thousands of emails from the past six years it plans to apply his approach to handling radio station listener contests, which vary by state and are complex.
Demonstrating the value of AI tools is not easy. Only one-fifth of companies using AI are collecting metrics to assess their return on investment (ROI), a recent Thomson Reuters survey showed. And those seeking to quantify the impact tend to focus on efficiency rather than how AI is helping to accelerate growth, retain skilled workers or reduce risk.
One thing is clear: amid increased competition from AI start-ups, legal tech companies need to strengthen the case for their products. "They're going to continue to push all of us to be better," McDonnell says.
.jpg%3Fw%3D1200%26ar%3D40%253A21%26auto%3Dformat%252Ccompress%26ogImage%3Dtrue%26mode%3Dcrop%26enlarge%3Dtrue%26overlay%3Dfalse%26overlay_position%3Dbottom%26overlay_width%3D100&w=1280&q=75){kind=small}
