Big AI's Anthropic edges into legal tech territory

Anthropic has released a new AI plugin for its enterprise product Claude Cowork, which can automate certain types of legal work.

The company says the plugin “automates contract review, NDA triage, compliance workflows, legal briefings, and templated responses - all configurable to your organisation's specific playbook and risk tolerances”.

Investors reacted quickly to the news with shares falling by 15% on Tuesday for legal data giants such as RELX, owner of LexisNexis, and Thomson Reuters, reflecting concerns that large AI developers could erode the value of more specialist legal research and competing workflow tools.

“Anthropic launched new capabilities for its Cowork product to the legal space, heightening competition within the space,” Morgan Stanley analysts said in a note on Thomson Reuters. “We view this as a sign of intensifying competition”.

Fresh competition

Until recently, most large AI providers positioned themselves as general productivity players, leaving legal-specific applications to vendors such as Harvey, Legora and other legal AI startups. Those tools typically sit on top of general-purpose AI models like ChatGPT or Claude, adding legal workflows, playbooks and safeguards designed for law firms and in-house teams.

Anthropic’s legal plugin suggests a shift in that balance. While the company has stopped short of launching a standalone legal product, the functionality described mirrors many of the features offered by established legal AI platforms, particularly around first-pass contract review and internal legal workflows.  

A report published last year found that 66% of in-house legal teams use general-purpose tools such as ChatGPT, while only around 17% are using dedicated legal AI tools. The findings suggested that ease of access and flexibility are often winning out over legal-specific tooling for many teams.

A familiar move from Big Tech

The move echoes developments from OpenAI, which has also moved closer to legal workflows over the past year. Updates to ChatGPT Enterprise have added enterprise-grade compliance controls, document connectors and agent-style features that support multi-step analysis.

Like Anthropic, OpenAI has thus far framed these capabilities as general productivity tools rather than legal advice engines. But in practice, these releases have made ChatGPT easier for law firms and in-house teams to deploy for routine legal tasks, in ways that overlap with specialist legal AI products.