Future of UX · 2026-06-25 · 27 min
Most teams default to chat when they think about adding AI to a product. But chat is just one pattern - and often not the right one. In this episode, Patricia breaks down the four AI interaction patterns every designer needs to know: Advisor, Generator, Agent, and Orchestrator. Each one has a completely different UX logic, a different failure mode, and a different answer to the one question that matters most: who carries the risk when the AI is wrong? Key Learnings: There are four distinct ways AI can show up in a product - and most teams only ever reach for one Each pattern has a different design job: skimmable, editable, reversible, legible The question "who carries the risk when the AI is wrong?" will change how you approach every AI feature you design Most real products are a mix of patterns - knowing which part is which is where the real design work starts Grab your early-bird ticket for the Future Product Days: AI for Designers : 5-week Bootcamp opening September →
#163 The Quarterly: My 24h Work week and What I'm Actually Working On (And What I Got Wrong)61 / 100
#160 AI Customer Research: Why Designers Should Learn Claude Code with Caitlin Sullivan63 / 100
#162 The Case of the Vanished Judgment#161 700 Components in One Day: How Design Work Is Quietly Moving Into Code#158 Why Junior Design Jobs Are Vanishing (And What To Do About It)