I was brought on to redesign appointment scheduling for a legacy EHR. A few weeks in, it was clear scheduling wasn't the actual problem. It was the symptom.
The platform had been built feature by feature, with engineering driving most of the decisions and no one really owning the product experience as a whole. Scheduling touched almost everything: front desk check-in, nurse review, physician treatment, post-visit documentation, billing. Fixing the calendar UI in isolation wouldn't have fixed what was actually broken underneath it.
So I pushed to reframe the brief. Instead of a scheduling redesign, we scoped a phased modernization plan across three areas (scheduling, patient demographics, and dashboard interactions) built around reusable patterns instead of one-off screens. That meant getting Product and Engineering realigned on a bigger scope than what had been asked for, and stepping into an informal lead role to hold that direction together.
Scheduling was the first piece we shipped, and it's where the reasoning was most concrete. The appointment queue was visually noisy, and clinicians were misreading appointment types at a glance, which meant more clicks just to confirm what they were looking at. We built a color system by appointment type, then added text labels directly on the tags rather than relying on color alone, since color-only encoding is easy to misread fast and easy to associate wrong. Icons handled the second layer, recurring versus telehealth, so the type and format of the visit were both legible without opening anything.
Average clicks to complete a scheduling task dropped from over 10 to under 5. CSAT with our beta adopters went from 55% to 74%. Time spent entering patient info also dropped, though we didn't have a clean way to quantify that one.
The part I'm most proud of isn't the color system. It's that the original brief would have solved a smaller problem than the one that was actually there, and it took pushing to get the scope right before the design work could matter.