AI agents are different. They don’t follow predictable paths, they act autonomously, and they fail in ways traditional software doesn’t. Most UX designers are still using familiar patterns for unfamiliar problems.

This workshop focuses on the hardest part: designing interfaces when you can’t control what happens next. Through a collaborative exercise, you’ll design a multi-agent system while working within real constraints like latency, unpredictable outputs, coordination between agents, and inevitable failures.

You’ll build frameworks for managing expectations, choosing what to show and what to hide, designing moments when users should intervene, and communicating uncertainty without breaking trust. The goal is a set of patterns you can use Monday morning: how to handle delegation, monitoring, handoffs, and errors.

This is for designers moving into AI work who want to understand both user needs and engineering reality. No technical background needed, just curiosity about systems that change as you use them.

This pre-conference course is included in the Bronze / Silver / Gold ticket bundles
or available to purchase as a standalone course.