UX for AI products
Designing AI that experts will actually trust
Explainability, confidence cues, evidence trails and a graceful way to disagree with the machine — the patterns that decide whether a careful professional adopts an AI feature or quietly routes around it.
Increasingly these are also legal obligations. The EU AI Act requires disclosure that a user is interacting with AI, meaningful human oversight of high-risk systems, and explanations a person can actually act on — alongside GDPR's right to an explanation for automated decisions. I prefer to design those obligations into the interface itself, rather than bolting on a disclaimer.
Confidence & uncertainty
Source grounding
Human approval points
Conversational UX
Agentic workflows
EU AI Act & GDPR
AI in the design practice
Faster process, same rigour
Research synthesis in hours, working prototypes in a day, audit agents watching accessibility and design-system conformance on every build — with the decisions still made by people.
Research synthesis
Rapid prototyping
Audit agents in CI
Data governance
Team adoption