Complexity Made Clear
Complexity Made Clear
Embedded Files
I design enterprise software for healthcare, where a confusing screen can turn into a compliance problem and a bad assumption can reach somebody's patient. Most of what lands on my desk didn't start on the screen. Whatever point a project is at when it reaches me, my job is the same: bring clarity, hold the focus, and fight for what users actually need. Do that well and the idea that started upstream becomes a rough prototype, then a real product, and the people using it are better off for it. AI helps me get there faster than I used to.
What AI actually changed
What AI actually changed
Without research, a polished interface is still a well-dressed guess. AI does not rescue shallow thinking. It scales it. What it genuinely does is clear away the mechanical work around research, synthesis, and documentation, which buys back the hours that should have gone to understanding the problem and testing the answer. Used without discipline, it turns an unexamined assumption into a production-ready mistake faster than any tool in the history of this field.
Synthesize Evidence
Synthesize Evidence
An LLM is the best research assistant I've ever had and the worst researcher I've ever met. It will organize six months of scattered findings in minutes, reconcile perspectives that would have taken me a week, and catch the contradiction I'd have walked past at four on a Friday. It will also invent a pattern that isn't there if I let it. So I make it show its work, and I go back to the source material myself.
Shape the Interface
Shape the Interface
Figma Make turns a validated workflow into something clickable fast enough that I can put a prototype in front of a user the same week the research wrapped. Speed is the point, but only because it buys more rounds of testing, not fewer. I still measure every direction it produces against the design system, the accessibility requirements, and the real constraints engineering already gave me.
HOW I WORK
Chaos into one source of truth
Chaos into one source of truth
Every project starts the same way. I pull the scattered inputs, the documenation, the feedback, the half-finished decisions, into tone govered palce, the work in an order that refuse to skep a single process.
Why any of this matters
Why any of this matters
Most of the AI use I've seen inside product teams is one prompt, one output, and a lot of hope. I've done that too, and it produced work I couldn't defend when someone asked a hard question. The process above is what I built so I'd stop having to.
Page updated
Google Sites
Report abuse