Interaction design
AI surfaces happy paths fast. Edge cases are still on you.
How AI is impacting this skill
Generative tools can spit out happy-path flows in minutes. The real work — error states, recovery, empty states, ambiguous inputs, agentic uncertainty — still needs a human who understands systems.
The shifts that matter
- AI-generated UI handles 80% scenarios but misses the messy 20%.
- Designing for AI features (uncertainty, hallucination, undo) is a new core skill.
- Flows now include agent actions, not just user actions.
- Permission, transparency, and trust are first-class interaction concerns.
What to do about it
- Document states, errors, and recovery for every flow you ship.
- Practice designing AI-touched flows: confidence, citations, undo, escalation.
- Audit one existing flow for failure modes and propose fixes.
- Build a personal checklist of edge cases you always cover.
Tools to try
Quickly explore interaction variants and states.
Prototype conditional and AI-driven logic.
Pressure-test flows with working code, not just visuals.
Next move
See how the rest of your skills stack up
Take the 3-minute UX skill assessment and get a focused roadmap.
Other skills
Synthesis is cheap. Good questions and good evidence aren't.
Framing the right problem matters more than ever.
Taste and systems thinking are the durable edge.
Systems are the bridge between AI output and shipped product.
Decisions move faster. Alignment is the bottleneck.
Design intuition is necessary but no longer sufficient.
The skill that compounds everything else.