Our product is an EMR system with a UI that hasn't meaningfully changed in about 30 years. That's the kind of fact that sounds fine internally — the system works, clients are used to it — until you're comparing screenshots against the competition and it stops sounding fine. Larger companies solve this with headcount: dedicated frontend developers, dedicated UI/UX designers. We don't have that team. What we have instead is support, custom requests, and customer service that actually respond — real strengths, but strengths that don't show up the moment someone opens the software for the first time.
Why nobody had fixed it yet
Because the design was so far behind, the changes needed weren't subtle — even modest effort produced a visible jump. And still, years of technical debt and the daily grind of feature work kept crowding it out. The bigger issue underneath that: not every developer on the team is a frontend developer, or has design instincts to begin with. Most of the team is functionality-oriented — they can ship a feature that works correctly, which is its own real skill, but making it also look like it belongs in 2026 requires React, CSS, HTML fluency, and an eye for which piece of data actually deserves the user's attention. Not everyone has both halves of that.
The eye I had, the skills I didn't
I used to work as a frontend developer before this job. That gave me the eye — I know what a clean interface looks like, and I know how to think about surfacing the data that matters instead of burying it next to everything else. What I no longer had was the day-to-day technical fluency to build it myself, fast, at the level the product actually needed.
This isn't specific to the EMR rebuild — I've designed dev dashboards and a handful of other tools I genuinely think look good, and AI has consistently been the thing that bridges the gap between having a design idea and actually being able to present it, working and styled, instead of just describing it in a meeting. That gap used to be the bottleneck. It isn't anymore, and it's a meaningful part of why I've been able to be as productive as I have at this job.
AI closed that gap. I used it to guide and architect the implementation itself — reusable, well-structured styles instead of one-off overrides, a real CSS architecture instead of patchwork, built with scalability in mind so the next update doesn't start from zero. I folded ADA compliance into the same pass, applying the same rules behind the compliance agent I'd already built, since accessibility is a rating we were going to have to chase eventually anyway — doing it inside the UI rebuild instead of as a separate project later meant not doing the work twice.
The demo
I demoed the updates today in our UI/UX meeting. The senior dev's reaction was the kind you want: he said it should go out immediately, if it weren't for contract acceptance and the approval process standing in the way. That's the actual bottleneck now — not whether the work holds up, but the process around getting it in front of clients.
It's not lost on me that this follows the same shape as the MCP server work: a gap the company doesn't have dedicated headcount for, and AI closing enough of the distance that I can be the one who does it — using an eye for design I already had, on skills I didn't.
Update: a second senior dev saw the demo and asked for it on our actual internal site — see the follow-up on rebuilding that frontend in three days.