I am the Lead Product Designer for AI at PitchBook. I own the vision and experience direction for Navigator, our central AI platform for private market intelligence.
Before PitchBook I founded and ran Rainfall, a ten-person studio, for seven years, and earlier I was Design Director at Fantasy.
Most AI roadmaps are written along one axis: what the model can do next. That is the axis engineering can schedule, so it is the one that gets planned. It is not the axis that decides whether any of it gets used.
The one that does is proximity. How close the intelligence sits to the thing a person is actually looking at. A weak model inside the page beats a strong one in a separate tab, because the cost of the separate tab is paid on every single question. Which is why the alternative to a blank canvas is not a better blank canvas. It is the product you already shipped, converging on the person using it.
I work at both ends of that. At PitchBook I set direction that design teams across the company build against. Outside of work I design a consumer AI product hands-on, every week, which is where most of what I know about running a design system under AI-assisted production comes from.
PitchBook Navigator, where you set a goal and agents do the job rather than answer a question.
Honest confidence, reasoning a user can inspect before acting on it, and outputs that open with what a finding means instead of what it is. I wrote the ones our team follows.
Built by me, on our MCP server, to test ideas in working software rather than in mocks.
How a model's answer gets presented inside a data product people make decisions with.
I write about earned autonomy, proximity, and why AI has to be shown rather than claimed.
Five positions I design from
