Dynamic Last-Mile Routing

Dispatch still runs on manual calls and fixed routes. Build the AI that plans, sequences and reroutes deliveries as conditions change.

EMX is the logistics arm of the 7X Group — the delivery operator behind this challenge. What you build here has a direct pathway into EMX and the wider 7X ecosystem.

5 / 5slots taken

Dataset_Intelligent Last-Mile Planning & Dynamic Route Optimisation_Final.xlsx

Unlocks on event day

Track full — all 5 slots taken

The problem

Every morning someone decides which parcels go on which van, and in what order. Then the courier redoes half of it by hand before leaving the depot.

Those decisions run on experience. Nobody can prove the sequence is the best one available, that the tightest delivery promises are protected, or that the route is not quietly burning extra kilometres, fuel and tolls along the way.

Then the day starts and the plan stops being true. A customer is not home. Another reschedules. New parcels get added, some get cancelled, and traffic does what traffic does. The plan built at 7am is still the plan at 3pm, because nothing exists to rebuild it. 7X is looking for a dispatch engine that plans the day properly, then keeps re-planning it as the day changes.

What we're seeing

Operations teams group shipments into routes by hand. Couriers then spend around an hour of every nine-hour day resequencing their own stops before they even set off.

Between 15 and 20% of planned deliveries hit something unexpected: a failed attempt, a reschedule, a cancellation, a parcel added mid-shift. The route plan rarely changes in response.

Not every stop costs the same. A bank delivery takes about six minutes, a general delivery about four, and towers, malls, ports, airports and free zones add waiting time that planning does not account for. (Figures are illustrative, modelled on real operations.)

The numbers to move

60%
Reduce manual dispatch planning effort by 60% or more
15%
Reduce total kilometres travelled per shipment by 10–15%
10%
Improve courier productivity (deliveries per working hour) by 10%
30%
Improve on-time delivery performance by 30% points

What a winning build looks like

Build a prototype that demonstrates measurable improvement over a static planning baseline using the provided dataset. Your solution should show how AI can reduce manual effort in delivery planning, improve route efficiency, and respond to real-time operational changes — with clear reasoning behind the decisions it makes.