Core concepts

Dispatch scoring

The factors, the weights, and the audit trail — why every assignment can answer “why this driver?”

10 min read · updated

The scoring model

When a booking needs a driver, the engine scores every eligible candidate. The score is a weighted sum of explicit factors — there is no black box between a booking and an assignment. Factors and weights are visible on the assignment record itself.

  • ETA: predicted time to pickup, corrected by learned traffic history.
  • Rating: rolling passenger feedback, recency-weighted.
  • Utilization: fairness pressure toward drivers with less recent work.
  • Zone policy: rank queues, airport rules, and territory constraints.

Weights are policy, and policy is yours

Operators tune factor weights per service level: an executive transfer can favor rating heavily while a busy taxi night favors ETA. Weight changes are versioned, and you can simulate a proposed weight set against historical bookings to see which assignments would have changed before committing.

Offers, timeouts, and cascades

The top-ranked driver receives the offer with a countdown. Declines and timeouts cascade down the ranking, each attempt logged. Dispatchers can override at any point — and the override is recorded with the same fidelity as the automation it replaced.

The audit answer

Every assignment stores its full scoring table: who was eligible, each factor value, the final ranking, and what happened to each offer. Driver disputes about job distribution end with a report, not an argument.