AI for UX/UI & Product Design

• Research questions, desk and competitor research
• Interview preparation, discussion guides and usability-test planning
• Working with transcripts, coding qualitative data and extracting themes
• Quotes, evidence and source traceability
• Separating participant evidence from AI interpretation

Exercise: analyse a small research dataset and compare AI-supported synthesis with a manual review.

• Problem framing, jobs to be done, assumptions and hypotheses
• Journey mapping, information architecture and user stories
• Flows, alternate routes, edge cases and error recovery
• Accessibility prompts
• Using AI as a critic rather than an answer machine

Exercise: generate several flow approaches for a product problem, then evaluate them against explicit user and business criteria.

• From UX requirements to UI, interface concepts and components
• Microcopy, empty states, error messages and onboarding
• Design-system consistency and generative UI
• Image generation where appropriate
• Reviewing visual hierarchy, the danger of polished nonsense, and inclusive design

Exercise: develop a complete screen set including normal, loading, empty, success and failure states, not just a happy-path mockup.

• Prompt-to-prototype workflows and Figma to interactive prototype
• Using AI coding tools for design validation
• Creating realistic data and connecting screens
• Testing alternate ideas quickly and choosing the right fidelity
• When generated code is useful to designers

Exercise: turn an interface concept into an interactive prototype that can actually be tested.

A button that triggers an AI model is not an AI experience strategy.

• Setting expectations, AI input design and suggested actions
• Streaming and waiting states, sources and citations
• Confidence and uncertainty, editing, correction and regeneration
• Memory, user control and permissions
• Confirmation before high-impact actions, failure and fallback, trust calibration

Final project: an AI-assisted case study showing your starting evidence, where AI was used and deliberately not used, synthesis, UX decisions, flow, interface states, an interactive prototype, and what changed because of testing.

AI can generate twenty flows before lunch. It cannot decide which one is right unless the criteria, evidence and context exist. That is why this course puts as much emphasis on critique as generation: the better you explain the problem and judge trade-offs, the more useful AI becomes.

AI tools change quickly. We review the teaching stack before each cohort; it typically includes ChatGPT, Claude, Gemini, Figma’s AI features, prototyping and build tools, and current research assistants.