Routing that accounts for the road as it is.

An ETA model trained on map speed limits will be wrong on every Indian route that matters. The useful signal is in your own historical trip data, which most fleets are already collecting and nobody is using.

01 / What is hard here

The problem underneath the problem.

Commercial routing here has to absorb variability that generic map services do not model: state border checks, restricted hours in city cores, monsoon conditions, road quality that varies by season, and driver rest requirements.

The other difficulty is that fleet data is honest about things people would rather it were not — idle time, route deviation, fuel discrepancy. Deploying these systems is as much a change-management problem as a technical one, and pretending otherwise is how they get quietly sabotaged.

02 / What we build

Work we take on in transportation.

ETA prediction: Arrival estimation trained on your own completed trips, which captures the corridor-specific reality that map-based estimates miss entirely.
Route and load optimisation: Multi-stop routing under vehicle capacity, time windows, driver hours and access restrictions — producing a plan a dispatcher can adjust.
Fleet utilisation: Idle time, deadhead running and asset utilisation analysis, with the actual constraint identified rather than assumed.
Fuel and maintenance analytics: Consumption modelling per vehicle, route and driver to separate genuine operating variation from loss.
Demand forecasting: Shipment and passenger volume forecasting by lane and time to support fleet sizing and positioning.

03 / Constraints

What shapes the build.

These are design constraints agreed at the start, not compliance work bolted on before launch. Retrofitting them is expensive and usually incomplete.

Telematics quality: GPS gaps, tampering and inconsistent device fitment are normal. Cleaning and validating this is a real phase, not a preprocessing step.
Driver acceptance: Systems perceived purely as surveillance get defeated. What is measured, what is shared and what it is used for should be settled and communicated openly.
Offline operation: Vehicles lose connectivity for long stretches. Anything on-vehicle must work offline and reconcile later.
Regulatory hours: Driver working-time rules are a hard constraint in the optimiser, not a preference to be traded away for a shorter route.

04 / What we advise against

Where we would say no.

We would advise against buying route optimisation before trip history is trustworthy, and against deploying driver-scoring without agreeing first how scores will and will not be used in employment decisions.

We would rather lose the work than build something that creates a liability for you or a harm for the people on the other side of it. If that rules out what you had in mind, it is better established now than at delivery.

05 / Other sectors

Adjacent work.

Different domains, the same delivery discipline — a measured baseline, a real evaluation set, and monitoring that catches decay early.

Next step

Working in transportation?

Tell us which process is slow, expensive or error-prone. The diagnostic will tell you whether this is a problem worth solving with machine learning — including if it is not.

Request a diagnostic All industries