Live fleet visibility and routing that holds delivery promises
Real-time fleet visibility, an optimizerthat sequences drops by timewindow, and a customer trackingportal for a last-mile operator.
The Challenge
Dozens of vehicles were coordinated entirely by phone, with dispatchers holding the plan in their heads and guessing at locations between check-in calls. Customers phoned in to ask where their delivery was, and the honest answer was that no one could see it; when a drop was disputed, there was no timestamp or photo to settle it. Fuel and overtime kept climbing, with no data to explain why.
Dispatch flew blind
Dozens of vehicles were coordinated by phone. Dispatchers guessed at locations between check-in calls, so one closed road could unravel an afternoon before anyone noticed.
No proof, only memory
When a customer disputed a delivery, there was no timestamp, photo, or signature to settle it. The record was a driver's word against a customer's recollection.
Costs drifting in the dark
Fuel and overtime climbed quarter over quarter, but with no trail of where vehicles actually went, there was no way to separate real demand from avoidable back-tracking.
Our approach
Every build follows the same software development life cycle, from requirements and design through build, testing, and support. Each phase is planned, demoed, and signed off before the next begins, so quality is engineered in rather than checked at the end.
Discovery & requirements
We fixed scope across tracking, dispatch, routing, and proof of delivery in a written specification, mapping the telematics, mapping, and traffic data the build depended on before any code.
(Outcome):
Architecture & design
We designed a real-time backbone and a routing core: a publish/subscribe streaming pipeline for live positions, a constraint-based optimizer, and a separate model for arrival times.
(Outcome):
Build
We built the live map, the driver app, the customer portal, and the routing engine in milestones, demoing the operation as each piece came online.
(Outcome):
Testing & UAT
We proved the optimizer on the operator's own history, backtesting against months of routes and replaying real delivery days in simulation, before it touched a live dispatch.
(Outcome):
Deployment & support
We rolled out in stages across the fleet, with drift monitoring on the ETA model and a dispatcher override that keeps a human in charge of the road.
(Outcome):
Outcomes
Around 40% fewer dispatch and 'where is my order' phone calls
On-time delivery up roughly 22% after routing went live
About 15% saved on routes and fuel in the first quarter
Timestamped photo or signature proof on every drop, so delivery disputes settle from the record