Stellantis x University of Michigan
AI-Powered Predictive Maintenance Ecosystem for the 2030 Jeep Gladiator
Role
UX Designer (team of 3)
Duration
4 months (Aug - Dec 2025)
Tools
Figma, Miro, Screens Studio, Sketchfab
Overview
Stellantis challenged junior designers at the University of Michigan to explore how AI could be integrated into vehicle systems to prevent unexpected breakdowns before they happen. We were given three prompts to choose from: AI-powered maintenance, EV battery optimization, and navigation. We chose maintenance immediately, and we chose the Gladiator just as quickly. Its reputation gave us a real, tangible problem to solve.
My role
I was one of three designers on this project, responsible for the AR mobile experience and companion app design, as well as contributing to research, comparative analysis, and the overall design system.
The problem
The Jeep Gladiator has a well-documented reputation for breaking down. Jeep drivers are a specific kind of person: adventurous, self-reliant, and deeply dependent on their vehicles whether they're on a mountain trail or a Tuesday morning commute. When something goes wrong, it doesn't just cost money. It breaks trust with a vehicle they rely on completely.
Research & discovery
I came into this project as a lifelong sports car enthusiast. My instinct was to design something refined and performance-driven. But the Gladiator driver is not an Alfa Romeo driver. They want rugged, dependable, and capable. I had to consciously set my own taste aside and design for someone else's life entirely. That shift in perspective was one of the most important things I did on this project.
Before touching Figma, I spent two months studying automotive display standards, safety regulations, and real in-car interfaces. I analyzed displays across five midsize crew-cab trucks: Rivian, Kia, Audi, Volvo, and Lexus. The weakest interfaces shared the same problems: outdated visuals, icons that required guesswork, and features that existed either on screen or as a physical button but never both. That redundancy gap influenced our own design directly.
The Pivot
Midway through the project, our research surfaced a finding we hadn't anticipated: the DIY automotive repair market has grown nearly 6% annually since 2020, and drivers who engage with it save an average of $3,993 per year. Our users weren't just open to handling maintenance themselves. Many of them actively wanted to.
Nobody was designing for that.
That single finding expanded our scope from one feature to three. It's also what gave us JeepView, the AR scanner I led the design of. The original brief asked for a maintenance page. The research told us drivers needed a teacher.
Challenges & Setbacks
The feedback that changed everything
Our original design displayed vehicle health information prominently on the home screen at all times
The Stellantis UX lead challenged us: "If a driver is constantly presented this information, will they just tune it out? Drivers do two main things in their car: navigate and listen to music."
We redesigned the widget to be scrollable and opt-in, surfacing health information only when the driver chooses to see it
The insight I didn't see coming
During a one-on-one user testing session, a user paused and said quietly: "What if my car health is low and a passenger sees that? That's kind of embarrassing."
I had been so focused on visibility and utility that I hadn't considered the social dimension of the interface at all
That one comment reinforced the opt-in direction and sent me back to rethink the entire home screen approach
1. Smart Display Integration
An intelligent notification system built into the center display and driver cluster. A swipeable home screen widget gives drivers the option to view vehicle health, but never forces it on them. Color-coded indicators (green, yellow, red) communicate severity instantly. Critical alerts surface in the driver cluster only when immediate attention is needed. The system stays quiet until it has something worth saying.
2. JeepView AR Scanner
A mobile app feature where drivers point their phone camera at any part of their vehicle. The AI identifies the part, explains its function, flags potential issues, and serves how-to repair videos when needed. It transforms maintenance from intimidating to something a driver can actually handle themselves.
3. Integrated Scheduling
Once a driver understands what needs fixing, the app suggests service windows based on issue severity and their existing calendar. Appointments sync natively with phone notifications and app reminders so nothing slips through.
Color
E3E3E3
Color
777777
Color
000000
Color
0280F9
Color
E40907
Color
FF9F10
Aa
Icons
Name
Weight
Size
Line Height
Heading 1
Heading 2
Body
Button
Bold
Bold
Regular
Bold
40
32
24
20
52
46
36
32
Impact
When we presented to the Stellantis UX team, the reaction that stuck with me most was around JeepView. The UX lead said it was genuinely innovative, something the team hadn't considered before. She told us this was exactly why Stellantis creates student projects like this: to discover thoughtful applications of emerging technology that internal teams don't always see.
Being engaged with critically by an industry professional, not just politely, was something I didn't fully anticipate as an undergraduate designer. She pushed back where it needed pushing and responded to our work like it was real. Because to her, it was.
What I learned
Designing within unfamiliar constraints requires patience before creativity.
Two months of learning before a single wireframe felt slow, but it was the only way to design something worth showing.
The most valuable feedback rarely comes from the most expected place.
A quiet comment from one user in a testing session reshaped the design more than any formal review did.
Good design means setting aside your own taste entirely.
The moment I stopped designing for the car I wanted and started designing for the person who actually drives this one, everything clicked.
Scope expansion is an opportunity, not a setback.
When our research uncovered the DIY maintenance trend mid-project, we could have stayed in our lane. Leaning into it gave us our most innovative feature.