Where the diesel went, on which leg, and who was driving
A Zimbabwean haulage operator running long-distance and cross-border loads knew diesel was leaking out of the business, but could never say on which trip, at which stop, or under which driver. This deployment joined the trackers, tank sensors, fuel bay camera and fuel card statements already in place into a reconciliation that closes leg by leg, with a driver record built from dated evidence instead of impressions.
The operational challenge
Diesel was the largest controllable cost in the business and the least controlled. Fuel was drawn at the depot bowser, topped up on a corridor account across the border and sometimes bought for cash, while consumption was judged on a monthly average that buried every bad leg inside a good one. Nothing tied a fuel event to a place, a time and a tank level, so the standing explanations — heavy load, bad road, days queuing at the border — could never be separated from theft. Driver risk sat in the same fog: the yard knew who was hard on equipment, but the workshop invoice arrived far too late to act on.
What the operator already had
A tracking contract on every horse, from a supplier the operator was reasonably happy with, plus separate units on high-value trailers after a hijacking. Newer trucks carried engine data on the CAN bus — the vehicle's internal network — which reports fuel used and engine hours far more honestly than a monthly average. The depot had cameras over the fuel bay and the gate, recording for security and reviewed only after an incident, and finance received fuel card statements as spreadsheets.
What we built
A fuel and driver layer above the existing tracking platform, reading positions and ignition through its API — the feed that lets one system pull records from another — rather than competing with it. Tank probes went in only where the fuel signal was unusable, and computer vision was pointed at the fuel bay camera already installed, so each depot dispense is matched to the vehicle actually at the pump. Every refuelling event is matched to a trip, a driver and a rise in tank level, and whatever will not reconcile becomes an exception carrying its evidence: the tank curve, the position at that moment, the card line and, at the depot, the clip.
How it works day to day
The transport manager opens a queue already sorted for him: unreconciled fuel first, then route deviations, then driver events from the previous shift. Each item carries its own evidence, so the conversation with a driver is about one stop at one place on one night rather than a general accusation at month end. What cannot wait — a tank level falling while the vehicle stands with the engine off, a trailer moving with no trip assigned — goes to WhatsApp and the duty phone.
Driver risk a driver can argue with
Trips are scored for the behaviour that predicts cost: harsh braking, speeding judged against the limit for that road rather than one fleet-wide figure, extended idling, night running and stops that were never on the route plan. Every score decomposes into named, dated events traceable to a tracker record or a camera clip, and each driver sees the view his supervisor sees. The record therefore exists before an incident, rather than being assembled afterwards to support a conclusion.
Integration points
Fuel exceptions post into the accounting system as coded transactions, so finance reconciles once instead of rebuilding the fuel position by hand. Trip records and driver hours feed HR and payroll automation where trip allowances are paid, removing the paper claim form. Load status flows to InOne CRM so a clerk can say where a load is without phoning a driver on the road. Where a weighbridge or gate access system exists, its ticket, the gate camera and the tracker arrival tie to the same trip record.
What it covers
Reconciliation per leg
Fuel drawn is matched to distance, load and engine data leg by leg, so a bad trip is no longer averaged away inside a good month.
The tank curve, not a snapshot
Fuel level is read as a shape over time, which is how a slow overnight siphon separates itself from normal consumption or a sloshing tank on a rough road.
Vision on the depot bowser
The existing fuel bay camera identifies which vehicle is actually being filled, so fuel booked against one registration and delivered into another becomes visible.
Border dwell counted separately
Queue time at a border post is tracked as its own state with its own idling profile, so it is never mistaken for an unauthorised stop.
Risk follows the driver
Events are recorded against the person on shift rather than the vehicle, so a driver's record travels with him between horses.
Your tracking contract stays
We read from the telematics supplier already in place, and add hardware only where a signal genuinely does not exist.
How it works
Site assessment
Week zero was spent in the depot and the fuel bay, not a boardroom: how fuel is issued, who signs, how trips are planned, which trucks report CAN data and which tracker units still work.
Integration and observation
We connected the tracking platform, fuel card statements and issuing records, fitted probes where needed and ran vision on the fuel bay. The system only watched at first, building a baseline of real consumption before it could accuse anybody of anything.
Rules and guardrails
Variance tolerances, route corridors, permitted stops, night movement rules and escalation paths were agreed with the transport manager and tested on historical trips, so alerting was calibrated on this fleet's own roads rather than a template.
Go live and optimise
Exceptions began reaching the duty phone and the morning queue, reviewed weekly so noisy rules could be retired and generous ones tightened. A first system of this kind runs approximately four to six weeks from assessment to go live.
Questions buyers ask
Do we have to leave our current tracking supplier?
No, and we usually advise against it at the start. If the units report reliably we integrate through their platform and add the fuel, trip and driver layer on top. Changing supplier stays a separate commercial decision.
Can drivers defeat the fuel sensors?
Any sensor can be interfered with and we will not claim otherwise. What changes is that interference stops being invisible: a probe that goes quiet, or a tank reading that no longer agrees with CAN consumption, becomes an exception in its own right. The strength is in cross-checks that must be defeated together.
Our trucks run where there is no coverage. Does it still work?
Units hold their records and upload when the connection returns, stamped with when the event happened rather than when it arrived, so a dead stretch fills in later instead of being lost. Live alerting pauses while a vehicle is out of reach, and the screen shows when each unit was last heard from.
Is this proof we can take to a disciplinary hearing?
It is strong operational evidence: timed, located and corroborated across independent sources, captured before anyone knew it would matter. Whether it meets the standard for a particular hearing or insurance claim is for your legal advisers to say.
How do drivers usually react?
Best where the record is open. Careful drivers finally have something in writing on their side, and arguments about fuel drawn or hours worked are settled from a record. Deployments go badly when a system appears unannounced, so tell the crew what is measured and show them their own view.
Where does our fleet and driver data live?
In your environment or a dedicated tenant in your name, with a retention period agreed in writing before go live. That includes movement and behaviour data about staff, scoped to the operational purpose you define.
Book a platform briefing
Fifteen minutes on your fleet: the trackers you run, how fuel is issued at the depot, and where your reconciliation breaks today. We will say plainly what can be built on what you already own. WhatsApp +263 77 636 6999 or sales@eigenstatesystems.com.