Future of UX | Your Design, Tech and User Experience Podcast | AI Design #149 Designing AI Features: The 3-Step Framework Product Teams Need with Bansi Mehta
37 snips
Mar 26, 2026 Bansi Mehta, founder of Koru UX Design and healthcare-focused UX leader, shares her Sense–Shape–Steer framework for designing AI features. She breaks down how to spot real AI opportunities. She compares copilots versus agentic experiences. She highlights storyboarding AI behavior and rapid prototyping to test ideas quickly.
AI Snips
Chapters
Transcript
Episode notes
Why AI Projects Stall On Arrival
- AI feature requests often arrive as working POCs that teams can't interpret or productize.
- Bansi built Sense–Shape–Steer because teams repeatedly restarted when briefs arrived as opaque engineer prototypes or vague pressure to "add AI."
Start With Problems Not Models
- Start AI design with Sense: map real user problems and where AI intersects with value rather than chasing novelty.
- Audit data access, model availability, RAG needs and failure consequences early to spot feasible, valuable targets.
Always Bring Engineers Into Early Workshops
- Include product and engineering in early Sense workshops so teams can surface prior POCs, data constraints and technical tradeoffs.
- Let product or design drive the session, but keep engineers present to avoid impossible assumptions.