Product Design · Fleet & Telematics

Designing a fleet system after the last one failed

TL;DR

The client's last reporting tool failed because users couldn't find anything. I built a guided 4-step report flow instead of a long list. The proposal secured the contract and completed a 20-vehicle POC.
Role

UX Designer

Contract outcome

Signed

Tools

Figma · Claude Code

Under NDA.

Project details have been withheld or altered. Design work, reasoning, and screens shown are my own.

01 / Problem

The client wasn't starting fresh. They were starting after a failure.

Before this project, the fleet team did most of their work by hand. Using spreadsheets, phone calls, manual fuel and toll claims. Two problems pushed them to look for a new system:

No visibility.

There was no live map. Managers had no way to know where a vehicle, or a driver, actually was.

The old reporting tool didn't work.

Staff create and submit reports every week. Fuel summaries, driver scorecards, trip reports. It's not optional, it's part of the job.

That second problem mattered most. This wasn't a fresh start, whatever I designed had to fix the exact thing that made people stop using the last tool.

02 / Users

Same fleet, two different jobs

Two kinds of people needed this system, and they needed to see different things.

ROLE 01

Operator

Manages the whole fleet. Need the full picture for vehicles, drivers, and maintenance.

ROLE 02

Dispatcher

Needs to know what's ready to go out right now. Not the whole fleet, just today's work.

The real design question wasn't "what should this tool do." It was "who needs to see what" so each screen only showed what that person needed to act on.

03 / Pivot (Report Generation)

Fixing what the last tool got wrong

At first, I thought the problem was missing reports. So my plan was simple: make sure every report type is available.

But the old tool already had every report type

Over 25 of them, all in one long list. The problem was never "missing reports." It was that nobody could find the one they wanted in that list.

So instead, I built a simple 4-step guided flow.

Step 1

Search or pick a category.

Step 2

Choose the report's details.

Step 3

Pick the format and schedule.

Step 4

Review and generate.

One long list became a short, guided path.

Constraint

I didn't get to test this with real users as there wasn't time in the proposal stage. The design was validated with 6+ stakeholders at the client, and their approval helped win the contract. But stakeholder approval isn't the same as staff using it daily. That's the part I'd still want to check if the project resumes.

04 / Pivot (Visibility & Safety)

A map dot shows "where." It doesn't show "how."

The easy fix for "managers don't know where vehicles are" is a live map with GPS. That's where I started.

But a dot on a map only tells you where a vehicle is.

It doesn't tell you how it's being driven, or if it should even be there. The client's real concern was:

Harsh driving

Unsafe behaviour

Personal use of company vehicles

So I added AI camera alerts alongside the live map.

The system flags harsh driving, no seatbelt, and driver fatigue on top of speed and geofence alerts. The map shows where. The AI shows how, and whether that's okay. This also matched what the company is known for which is AI and tech. This made it a stronger pitch than GPS alone.

Screens

05 / Outcome

The proposal secured the contract. The project paused after POC.

The design work (report flow, live map, AI alert layer, role-scoped views) was enough to win client sign-off and move from proposal to commissioned project. A proof of concept ran across 20 vehicles and 3 branches.

Signed

Contract secured from proposal work

20 / 3

Vehicles · Branches in POC

Paused

Status after POC, reason unknown

The project didn't progress past this point, so I don't have production data. The design work secured the client.

06 / Reflection

What I'd want to know now

Stakeholders approved the design in a room. That's not the same as staff actually using it every day.

If the project starts again, here's where I'd start

The report tool

I'd want to test with real staff doing their actual job.

Does the 4-step guided flow actually help field staff find what they need faster, and does it hold up under the full range of report types they'd actually request?

Thank you for reading.