A heavy-haul fleet ran around the clock and still watched its route profitability slide. The reason was the dispatch desk: hundreds of trucks placed by phone calls, WhatsApp groups and gut feel — queues at the quarries, and half the trips driven home empty.
Hauling quarried material and earth is a game of tonnes and timing — every truck idling in a queue or driving home empty is burning fuel, wages and money.
The fleet ran around the clock, and route profitability kept eroding all the same. The root cause sat at the dispatch desk: hundreds of trucks positioned by hand, from guesswork, phone calls and WhatsApp groups. It bred logistical chaos — empty runs, and giant queues at the quarry gates.
Three failures compounded. Dozens of trucks hit the loading and weigh stations in the same morning peak, creating hours-long queues that idled expensive haulage capacity. A truck that tipped its load at a site in the north drove all the way back to a quarry in the south completely empty, because the dispatcher had no way to spot a pickup or available material on its return route. And urgent changes at the concrete plants and sites — a pour cancelled by weather — never synced back to the fleet, so trucks rolled up to sites that were closed or jammed.
One picture of the whole fleet, staggered timing, and a return load found automatically.
We connected the live GPS feeds from the trucks, the load sensors at the quarry weighbridges, and the customer-order system into one scheduling platform — so the desk finally saw the whole operation at once.
An optimisation model reads vehicle positions, traffic and quarry queues in real time and produces a staggered schedule — so trucks stop bunching at the same weigh station and loading capacity isn't wasted in a peak.
The system spots a truck that has finished tipping and automatically finds a return load in its area, pushing the dispatcher a one-tap assignment — turning an empty drive home into a paid one.
Changes at the plants and sites flow straight to the fleet, so a job called off by weather redirects the vehicle in the moment instead of stranding it at a closed gate.
Instead of chasing drivers by phone, the desk manages hundreds of trucks from one board — by exception, not by rumour — and spends its attention where it changes the day.
Nothing reassigns on its own. The AI proposes the route, the stagger and the backhaul; the dispatcher approves with a tap. Judgement stays with the person who knows the ground.
The trucks stopped waiting, and stopped driving home empty.
Average waiting time in the loading and weigh queues fell 30%, and empty return trips dropped 20% — a large monthly saving in fuel and in tyre-and-component wear across the fleet.
The company ran about 15% more loads and deliveries a day with the very same number of trucks and drivers — and hit its delivery windows far more reliably, lifting on-time-in-full performance for its end customers.
Nothing here bought a single new truck — the gains came from ending the empty miles and the queues that were quietly eating the fleet's own capacity. It's the same live-sync-plus-prediction discipline behind our multi-site schedule-control work, where seeing the real picture ahead of time is what lets a person rebalance before the jam. The dispatcher still approves every move; the AI just makes the good move obvious.
Reef TRH