About

AI agents sound like something that needs a development team, a budget line and a six-month project plan. They are not. The tools now exist for a legal leader to build a working agent, one that takes a task, works through it step by step and hands back a finished output, in the time it takes to drink a coffee.

In this live-build workshop, Ben Palmer, C-Suite Director at Winmark, constructs a working agent from scratch in front of the room: a first-pass NDA reviewer that checks an incoming contract against standard terms and flags every deviation, built in under twenty minutes on a free account, with each step explained as he goes. Then the room does it. Every participant builds their own agent to the same brief and tests it against a contract seeded with hidden problems, so each person finds out on the spot whether theirs works. Catherine Qi of Winmark's research team then brings the approach to life beyond the demo, sharing a working example of how the same techniques are already being used inside Winmark's own research operation. The session closes with an open discussion on what this means in practice: which legal tasks genuinely suit agents, where judgement must stay human, what escalation triggers and audit trails would look like, and how to move from a demo to something your team would actually trust.

Designed for General Counsel and legal leaders who want to understand agents by building one rather than reading about them. No coding, no technical background and no IT involvement required: everything runs on a free claude.ai account, and all materials are provided and entirely fictional, so no company documents are involved at any point. Participants leave having built their first working agent, knowing exactly what an agent is, what it is not, and what it would take to use one properly.