CASE STUDY THREE
Preventing a Costly Detour Through Rapid Research
Role: Product Designer
Tools: Figma Make, Chat GPT, HeyMarvin, Figma
Tools: Figma Make, Chat GPT, HeyMarvin, Figma
The Problem
Fleet managers are responsible for keeping hundreds of vehicles operational, but the information they need to make maintenance decisions lives across multiple parts of Fleetio.
Service reminders, vehicle issues, maintenance history, and upcoming work all exist, but connecting those signals requires manual effort. As one fleet manager described it: "It's three or four different places to find that information." The result was a fragmented decision-making process that relied heavily on experience, memory, and manual investigation.
For newer fleets, maintenance was often prioritized reactively. For more mature fleets, prioritization existed, but it was still largely manual.
The Opportunity
Our initial hypothesis was ambitious.
What if AI could analyze maintenance signals, prioritize work automatically, and tell fleet managers what needed attention first?
The concept became Maintenance Triage: a centralized workspace that combined maintenance signals with AI-driven prioritization and recommendations. Before committing engineering resources, we wanted to understand whether the problem, solution, and workflow actually resonated with customers.
The Approach
Using Figma Make, we rapidly built a functional prototype and conducted concept testing with fleet managers across multiple fleet sizes.
The entire research effort--from prototype creation to synthesis and recommendations--was completed in roughly 1.5 weeks. AI-assisted workflows accelerated transcription, synthesis, reporting, and prototype creation, allowing us to validate the concept before development began. Our goal wasn't to validate the solution.
It was to uncover what customers actually found valuable.
Learnings
The most important finding surprised us. Users consistently loved the consolidation. They were far less excited about the prioritization.
Across interviews, customers repeatedly responded to the idea of seeing issues, service reminders, and recommended next actions in a single place. The value wasn't that the system was deciding for them. The value was that it eliminated the need to hunt for information.
The real problem wasn't prioritization. It was fragmentation. The strongest skepticism emerged around AI-generated priorities.
Experienced fleet managers trusted their own judgment more than an algorithm, while newer fleets saw potential value in guidance. The feedback suggested that prioritization could become valuable over time, but only after trust had been established.
The Impact
The research fundamentally changed the direction of the project.
Rather than leading with AI prioritization, we shifted the product strategy toward consolidation: creating a single place to view maintenance signals and next steps. This approach delivered the highest immediate customer value, carried significantly lower adoption risk, and created a stronger foundation for introducing AI-assisted prioritization in the future.
What began as an AI prioritization project became a workflow simplification project. And that pivot happened before a single line of production code was written.
CASE STUDY THREE