The challenge
EvolveVision Logistics runs 450 trucks across South Africa and cross-border routes into Botswana and Zimbabwe. Dispatching happened over phone calls and a wall of whiteboards. Each depot kept its own spreadsheet, and head office only learned about a problem after it had already cost money.
The waste showed up everywhere. Trucks drove back empty because one depot could not see another depot’s loads. Customers phoned for delivery updates because nobody could answer without radioing the driver. And utilisation, the number the business lives on, was calculated once a month, three weeks too late to act on.
EvolveVision had telematics boxes in every truck. The data existed. It just went nowhere useful.
The approach
We spent the first three weeks in depots, not in meetings. We rode along on dispatch shifts in Johannesburg and Durban and watched where the real decisions got made. The plan that came out of it was deliberate: put live vehicle data in front of dispatchers first, because a platform they trust gets used, and one they do not gets a whiteboard next to it.
- Shadowed dispatch teams at three depots to map how loads actually get assigned
- Streamed the existing telematics feeds into one event pipeline instead of installing new hardware
- Rolled out depot by depot, starting with the team most eager to switch
- Ran the whiteboards and the platform side by side for one month per depot before cutover
- Designed for spotty connectivity, because cross-border routes lose signal for hours
What we built
The platform ingests position and engine data from every truck through a Kafka pipeline and turns it into a live operations picture. Dispatchers assign loads on a drag-and-drop board that already knows where every vehicle is, what it is carrying, and when the driver runs out of legal hours.
- A live fleet map and dispatch board covering all 450 vehicles across seven depots
- Load matching that flags empty return legs and suggests nearby freight
- Automated customer notifications with delivery windows, replacing phone-in queries
- Driver hours and compliance tracking with alerts before limits are breached
- Utilisation, cost-per-kilometre, and on-time dashboards updated hourly, not monthly
- A resilient cloud architecture on AWS that buffers telemetry through connectivity dead zones
7
depots running on one shared platform
85%
fewer inbound "where is my delivery" calls
30 sec
from truck event to dispatcher screen
The outcome
Six months after full rollout, fleet utilisation was up 30%. Empty kilometres dropped 19% because dispatchers finally saw loads across depot boundaries and matched return legs that used to run empty. EvolveVision moved more freight in the same quarter with the same fleet, which went straight to margin.
The operational change stuck because the platform earned trust rather than demanding it. Every depot cut over voluntarily once they had watched the numbers for a month. Head office now runs a weekly review off the same live data the dispatchers use, and the monthly spreadsheet ritual is gone.
