Extract from Trudy Knockless’s article “How In-House Teams Are Using AI Agents—Without Letting Risk Run Wild”
Legal departments are under pressure to move faster without losing control, and a growing number of in-house leaders say AI agents are starting to offer a workable answer—handling high-volume, multi-step work while lawyers retain judgment, privilege and risk oversight.
An AI agent, in this context, is software that can take a goal—like reviewing a contract or answering an internal request—and carry out a series of steps to complete it, often by pulling from multiple systems. That is different from the broader idea of “agentic AI,” which refers to more autonomous systems that can plan, coordinate and act with less human direction.
Most legal departments today are not handing over work to fully autonomous systems. Instead, they are deploying more limited agents embedded in workflows, with humans still in the loop.
That shift is no longer limited to drafting and research. Legal leaders described agents being used for triage, due diligence, contract workflows, internal-request management, compliance analysis and outside counsel support.
