Legal departments shift from testing AI to governing it as agentic tools expand

Corporate legal departments have moved beyond experimenting with generative AI to governing autonomous agentic systems now operating in production. This shift is evident in three key developments: general counsel are embedding AI-specific protections into vendor contracts to manage risks like legal noncompliance and third-party harms from autonomous decisions; legal AI provider Harvey has launched tools allowing law firms to build custom agents, serving over 700 customers and 70% of AmLaw 10 firms; and a survey by FTI Consulting and Relativity shows generative AI adoption in legal departments surged from 44% to 87% in one year, with formal technology roadmaps more than doubling to 53%. The most common use cases include summarization, contract clause identification, and e-discovery. For legal professionals, the priority is now competent management of these systems, including understanding AI addendum language, ensuring quality of AI-produced analysis, and adapting liability frameworks. Structured training and workflows will be essential to work effectively alongside autonomous agents that review contracts and flag risks without human instruction, marking the end of what David Horrigan called the "era of the Luddite Lawyer."

(Source：Complete Ai Training)