About

In 2016, I spent a year studying how a robot's voice changes whether people trust it. It felt like a niche question then. It is the entire industry now.

Since then I've scaled a design function from one person to twenty-two, on a platform used by more than 7,000 global brands [VERIFY] brand count + platform valuation. I've led design through three eras of the same product — from integration workflows, to AI copilots, to agents that do the work themselves. Each transition rewrote what "designing" means, and each time the answer was the same: get closer to the material. So I build. I prototype agent workspaces in code, ship internal tools that change how my team works, and keep a lab of experiments that don't ask permission.

The thread through all of it is trust. Enterprise AI doesn't fail because models are weak; it fails because people can't see what the system is doing, or why. I've designed the unglamorous machinery of trust — identity, permissions, admin consoles, audit trails — and I've studied its human side in the lab. My research on how voice design shapes trust in AI advisors is currently in review at the International Journal of Social Robotics, ten years after the thesis that started it [VERIFY] final publication status before launch.

I think in physical metaphors — murmurations, dominoes, magnetic tiles — because systems are physical to me: touch one piece and the whole reshapes. That's the hero of this site, and it's also how I run teams. Change what designers can build, and you change what they believe is possible.