Fleet Operations Platform
Logistics & Supply Chain
2025
enterprise
What they needed
The operator dispatches roughly 1,200 vehicles across a six-state distribution network, moving freight for manufacturers and retailers on tight delivery windows. Dispatchers tracked trucks by phone check-in and a shared spreadsheet, updated every two to three hours at best. Routes were planned the night before and rarely adjusted for traffic, weather, or last-minute order changes, so drivers idled in jams while faster paths sat unused. Fuel costs were split across client contracts using flat mileage estimates rather than actual routes driven, which triggered repeated billing disputes and quietly ate into margins on longer hauls. As the fleet grew past 1,000 vehicles, a six-person dispatch team couldn't keep pace, and on-time delivery had slipped to 81%.
What we built
Prixelo built a fleet operations platform centered on a live map: a React dashboard showing every truck's position, updated via WebSocket connections streaming GPS pings every 15 seconds from onboard telematics units. A Node.js service ingests that stream, matches it against Mapbox's routing and traffic data, and recalculates optimal routes on a rolling basis, rerouting drivers around congestion or closures without a dispatcher touching a map. PostgreSQL stores the full trip history: every waypoint, stop, and idle period, tied to the client shipment it served. That history feeds an automated fuel-cost allocation engine, which apportions each truck's actual fuel spend across shipments by distance and load weight instead of flat estimates, generating client-ready billing breakdowns dispatchers used to build by hand.
How we approached it
Chose WebSocket over polling for live position updates, so the dispatcher map reflects each new GPS ping the moment it arrives instead of on a delay.
Layered Mapbox's traffic-aware routing on top of Node.js recalculation jobs that re-score every active route every 60 seconds against live conditions.
Modeled fuel allocation in PostgreSQL as a ledger tied to trip segments, not vehicles, so costs split correctly when one truck served multiple client shipments per run.
Built the dispatcher UI in React around a single live map view rather than a status table, matching how dispatchers actually think about the fleet.
The impact
On-Time Delivery Rate
Fewer Fuel-Cost Billing Disputes
Dispatcher Hours Saved Weekly
"We used to find out about a stuck truck when the driver called us. Now we see it before the driver does, and the system's already found a way around it. Fuel billing arguments with clients are basically gone."
VP of Operations
Logistics operator
Tell us what
to build.
Bring us the one everyone said was too hard. Free, no-obligation reply within 24 hours — NDA on request.