Agentic Experience Design: A New Interaction Layer Demands a New Discipline
summary
Argues that adoption fails from poor experience design not model power, based on 18 months of Slack DevRel data and platform signals like >1B daily messages and sudden builder growth. Shows agents shift the loop by owning planning and outcomes while humans keep intent and judgment, so trust breaks unless designs force context as an accountability surface. Recommends three guardrails — transparent, steerable, resilient — plus progressive authority, bounded autonomy, and shipping one thing done well.
AI-written summary · about
chapters (9)
- 0:00Introduction and motivation
- 3:38Explosive growth of AI and builders
- 10:41History of HCI models
- 14:50Agentic systems as new layer
- 18:27Defining Agentic Experience Design
- 25:31Building trustworthy agent systems
- 26:56Reflections on industry and users
- 29:44Q&A: Analogy to interns and skills
- 31:31Q&A: Psychological safety in systems
