North Light
Getting seven stakeholders to agree on what they're actually building.
Strategy
Enterprise
Role
Product Designer
Timeline
8 weeks
team
2 Enginners, 1 PM, me
platform
Web
The Real Problem
Therapists at Meridian Health were documenting 3-5 client sessions daily. Each one meant clicking through 12 different screens, re-entering the same information multiple times, and hoping nothing crashed before they finished. People were losing work, getting disoriented, and spending 15-20 minutes per note when they had clients waiting.
But here's what made it worse: the interface gave no indication of progress. Therapists would finish what felt like the last screen, only to discover three more hidden steps. The 'back' button sometimes saved your work and sometimes didn't. There was no auto-save, so one accidental click could erase 15 minutes of careful documentation.
The feedback we got was pointed:
'I genuinely don't know if I'm done or if there are more screens hiding somewhere.'
'I've lost the same note three times this week.'
'Why am I entering the session date in four different places?'
The underlying issue wasn't just bad UI—it was that nobody had ever looked at the complete journey. Each screen had been built separately over time, and they'd never been stitched into a coherent experience.
Finding the Fix
I started by mapping the entire 12-screen flow—not just the UI, but how data moved between screens, where duplication happened, and where people got stuck. After analyzing support tickets and survey responses, three problems stood out:
No sense of progress. People couldn't tell where they were or what was left.
No safety net. One wrong click and your work vanished.
Clunky information architecture. Screens were organized by database structure, not by how therapists actually think about documentation.
The clinical directors I interviewed described session notes as 'telling the story of what happened'—context, observations, interventions, next steps. But the interface chopped that story into arbitrary fragments.
So I focused on three fixes:
A progress indicator that showed completed sections, current section, and what's remaining—always visible, always clear.
Auto-save with real feedback. Not just saving in the background, but showing therapists exactly when their work was protected.
Consolidating screens from 12 to 7 by grouping related information the way therapists actually think about it, not the way the database happened to be structured.
All achievable within our constraints. No backend overhaul required.
How does it work?
The Main Screens
What I'd Do Differently
I'd push harder for access to real users earlier. Working through intermediaries gave me valuable insights, but I missed nuances that only come from watching someone actually struggle with an interface. Even one shadowing session would have accelerated my understanding.
I'd also document existing problems more systematically. I relied heavily on anecdotal feedback and support tickets, which worked—but a proper heuristic evaluation would have given me clearer evidence when advocating for changes.
What I Learned
Strategic improvements beat perfect overhauls. I wanted to rebuild everything into a sleek single-page experience. But given our constraints, that wasn't realistic. The three focused changes we made—progress visibility, auto-save, better IA—delivered serious value without requiring a ground-up rebuild.
Mental models beat logic. What made sense to me (grouping related data) didn't always match how therapists thought. Validating assumptions with people who actually do the work saved me from shipping something technically correct but functionally wrong.
Invisible design builds trust. The 'Last saved' indicator wasn't technically necessary—the system was auto-saving either way. But it transformed how people felt about the experience. Sometimes the most important design work is making invisible processes visible.
This is sample content for portfolio development purposes. Replace with your actual case studies when ready.