Practical applications of LLMs and generative AI across the research and design lifecycle, compressing the path from raw signal to confident decision without losing the human in the loop.
AI has changed the economics of UX work. Tasks that used to consume days (transcript coding, competitor teardowns, first-pass concept exploration) can now happen in hours. That changes what's possible, but it doesn't change what's true. The craft is still in framing the right question, recognizing weak signal, and translating insight into design direction the team can actually act on.
I treat AI as a force multiplier on the practitioner. I use it heavily where it shortens the distance between data and decision, and I keep it out of the places where human judgment, lived context, or stakeholder trust are the actual deliverable.
AI does not replace talking to users. It does not replace the judgment of an experienced practitioner reading a room. It does not replace stakeholder alignment, the political work of shipping good design inside a real organization, or accountability for the recommendation. Anything AI produces is treated as a draft: cited, verified, and revised before it leaves my desk.
If you're building (or rebuilding) a research and design function that uses AI seriously, I'd love to talk.