AI is evolving beyond simple, predetermined chat responses—new agentic systems can now make decisions and take actions on their own. This creates fresh challenges for UX designers, especially in healthcare where AI might soon be scheduling appointments, flagging diagnostic results, or even recommending treatment adjustments without constant human oversight.

In this hands-on workshop, we’ll discuss the big questions using real examples from Microsoft’s AutoGen framework. AI can be fantastically resourceful in seeking solutions, but how might we design for AI in healthcare by setting limits on its choices? When should clinicians be able to step in and override the system? How might we build trust through safeguarding when AI is acting independently in medical situations? Through practical exercises with AutoGen scenarios, we’ll explore experiences where healthcare professionals and AI could begin to work together effectively, with clear boundaries about what the AI can and can’t do on its own.