UX: Designing for Copilot
summary
Claims a copilot is a collaborative UX built on a design system (Fluent 2) and that architecture choices—model selection, hosting, grounding data, plugins—directly change trust, accuracy and latency. Guidance includes an AI notice, prompt suggestions, explicit citations and a continuous feedback loop; Azure teams test with customers, use GPT‑4 to evaluate GPT‑3.5, and favor modest rollouts. Recommendations: invest in content design, guardrails and telemetry so the assistant’s outputs are steerable and auditable.
AI-written summary · about
chapters (8)
- 0:00Start
- 1:26What's a copilot?
- 3:57A New Collaborative UX
- 4:30Design principles
- 6:48appropriate trust
- 7:36System anatomy
- 8:35Evolving the design system
- 21:45In practice: Lessons from the field
related talks
46mRedefining User Experience with Large Language Models
42mUX+AI: Mastering the UX of AI-Powered Products
48mBuild your own copilot with Copilot Studio
45mBuilding a copilot: Azure AI Studio or Copilot Studio
14mon Collaboration, Copilot & the Future of Work
51mThe Best Way To Build An AI Copilot Product? Fusion Teams!
46mBuilding Copilots - key lessons and best practices
36mCustom copilot Teams Integration in Government applications
