The release of legal-focused plugins for Anthropic’s Claude Cowork sent a jolt through the legal-technology and publishing sector in early February 2026, wiping billions from the market value of established providers and reviving a broader question: whether general-purpose agentic AI could undercut or even replace software built for specific professions such as law, sales and finance.
What Anthropic launched
Claude Cowork is an agentic desktop application Anthropic introduced in January 2026. Unlike a conventional chatbot, it can plan, execute and iterate through multi-step workflows rather than simply answering individual questions, and it can be granted access to a folder on a user’s computer so it can read, edit and create files there. The new plugins let users tailor these agentic capabilities to particular industries. A sales plugin, for instance, can connect Claude to a team’s customer relationship management (CRM) system and follow a defined sales process, while a legal plugin can help legal teams review documents, flag risks and track compliance.
Why the market reacted
The announcement triggered sharp falls in the shares of legal and data-driven information companies. Thomson Reuters dropped by as much as 18 percent in some trading, and RELX, the parent company of LexisNexis, fell around 14 percent in what was reported as its steepest single-day decline in decades. Dutch provider Wolters Kluwer and the London Stock Exchange Group also slid. Investors appeared to be pricing in the risk that a capable, general-purpose agent able to perform document review and compliance tasks could erode demand for specialised legal-research and workflow products.
Practitioners echoed that concern. A Chicago law-firm owner argued that this represented genuine agentic AI built specifically for legal tasks rather than a general chatbot, and that Anthropic’s ability to build and ship such agents at scale made it a direct competitor to established legal-information providers, some of which offer their own AI assistants. An academic specialising in data-science and AI strategy suggested that experienced legal specialists would likely benefit from such tools in the short term, even as the longer-term effect on the profession remained uncertain.
What to watch
Dramatic market moves on the day of an AI announcement are not the same as proven disruption, and several observers cautioned that early reactions to emerging technology often overstate near-term impact; large single-day swings can also reverse quickly. The practical value of legal AI plugins depends on accuracy, and general-purpose models can produce plausible but wrong answers, which in a legal context carries real professional and ethical risk. Established providers hold advantages that a general agent does not automatically replicate, including licensed proprietary databases, citation-checked case law and deep domain integration, and many are shipping their own agents. The realistic near-term picture is likely one of augmentation, with professionals using such tools under supervision, rather than wholesale replacement. Whether the competitive threat proves as severe as the initial sell-off implied will become clear only as adoption, reliability and regulation develop. For wider context, see coverage of Anthropic’s scale and funding and the rise of dedicated AI leadership roles in enterprises. Independent reporting is available from Investing.com and AI Business.
