I like making complicated products easier to understand. My work sits between product strategy and execution: research to frame the problem, structure for complex information, interaction systems, quick prototypes, then close work with product and engineering to ship. Most of it has been in regulated, high stakes territory, insurance quoting and financial decision tools, where clarity and control decide whether people finish.
AI is part of how I work now, mostly for research synthesis, exploration, and faster prototyping. It has not replaced the judgment part. The conversational insurance flow is not an AI agent, and I do not pretend it is, but designing it taught me what these products need: one decision at a time, states people can read, answers they can edit, and a way back when something goes wrong. That is the same ground agentic products stand on.