CRM for Car Dealership. Nobody was told to switch to the mobile app.
The brief was to copy a legacy Windows app. Their own usage data showed it was solving the wrong problem, so I changed direction.
Context
A US car dealership network was moving its CRM off a legacy Windows app. The brief was one line: copy the old app exactly, same logic, same layout, same flows, reusing the existing database for backward compatibility. We delivered that first. Then I took some room to explore, and used it to gather the data that backed a second version, a mobile app. The users came with me.
I ran the project as designer and, in practice, project manager, leading a small team of developers and QA. The constraint was strict and the deadlines were tight, delivered in Agile sprints.
Copying the desktop made sense: existing dealers were fast on the old system, and nobody wanted to retrain them. But the business also wanted to bring in new dealers, and for them the old interface was a wall.
Investigation
Instead of assuming the migration was working, I checked. I tracked adoption in the database and saw that people were not moving to the new system. To understand why, I pulled the actual documents from one branch.
Over half were phone photos, not scans.
Dealers were photographing paperwork on site and closing deals in the field, not at a desk.
Reframe
That changed the problem. The issue was not the age of the app, it was context. The work had already left the office, and any faithful copy of a desktop tool would keep missing where the job actually happened.
Approach
So I redirected the work. We built a mobile, offline-first app designed around fieldwork rather than a desk. To ship one codebase across Android, iOS, and web, the team adopted Flutter and Dart, both new to us at the time.
I grew a design system alongside the product so people who were not designers could assemble new screens without breaking consistency. We scheduled refactors before technical debt slowed us down, ran retrospectives, and tested with real dealers throughout.
My role
I owned both the design and the product. That covered the brief and its contradiction, the scope, the information architecture, and the interface. I also led the developers and QA who built it.
Outcome
Within two to three months, over 60% of deals ran through mobile. Nobody was forced onto it. People switched because it worked better for them than the tool they already knew. It also outgrew the brief: the app was clear enough that the owner used it to onboard new dealers.
This project is under NDA. The overview is all I can share.
