I’ve been doing this for twenty years, and here’s the thing nobody tells you when you start: the hard part was never the interface. It was getting someone to trust something new enough to actually use it.

AI hasn’t changed that problem. It’s made it sharper. A model can be brilliant and still sit there unused, because the person on the other end doesn’t know why to believe it, or what they’re supposed to do with what it just told them.

So that’s where I live now; not in the model, but in the space right next to it. The onboarding moment where a new AI agent either earns someone’s trust or doesn’t. The dashboard that turns a wall of output into a decision someone can actually make. The quiet framework that helps a team stop guessing and start designing with AI on purpose, instead of bolting it on after the fact.

I’m not trying to make the AI smarter. I’m trying to make sure the human next to it isn’t left guessing.

Right now that means applied-AI product work, decision-support design, and slowly building a public case for what “good” looks like when people and AI are actually deciding things together, not just co-existing.