The After Visit Summary, shipped to 90M+ patients. You just scrolled the short version.
Design and engineering drift because they're two artifacts pretending to be one. I grounded my design system in the same component architecture engineers ship (shadcn/ui), so parity isn't maintained; it's built in. Claude handles mechanical exploration; I keep judgment, strategy, and craft.
Read the case study →When the org disbanded its consistency-review and accessibility teams, the checks disappeared with the headcount. I rebuilt them as Claude skills spanning the design lifecycle: self-check against the design system, automated handoff annotation, research on tap. The scaffolding returned. Judgment stayed human.
Read the case study →64% of traffic was mobile, but records read like bureaucracy on a four-inch screen. We rebuilt native around comprehension, not navigation: every number got a frame of reference, every summary led with what to do next. Accessibility wasn't a layer; it was the design.
Read the case study →The legacy kiosk demanded everything (payment, consents, screeners) before giving patients the one thing they wanted: certainty. I stripped check-in from 10+ steps to five, told patients exactly where they stood, and designed for the accessibility hardware before it existed. It became the company's kiosk standard.
Read the case study →I'm Paolo Conti, a product designer who works the full stack of the craft: strategy at ten thousand feet, systems at a thousand, pixels at one.
The last decade shipped healthcare experiences to 90+ million people: native records read by patients who'd stopped reading them, a kiosk that became a national standard. The last two years rebuilt how I work: an AI-native pipeline where design and code share one source of truth, and quality checks live inside the work instead of a review queue.
I do my best work on teams where that's the starting point.