Results Logistics and transport

Route planning for a fleet of 420 vehicles

A planning engine that recalculates routes in eight minutes instead of four hours of manual work, with 12% fewer kilometres driven.

−12% kilometres driven

8 min replanning time

96% on-time deliveries

The challenge

Daily planning was done by hand, by six dispatchers, over four hours. Any last-minute change — a cancelled order, a broken-down vehicle — triggered manual replanning of the whole day.

The application used by drivers did not work without a signal, and 30% of the areas served had poor coverage.

The solution

We built an optimisation engine that accounts for delivery windows, capacity, city access restrictions and legal driving times.

Full replanning in eight minutes, on any change
An Android application that works offline and syncs when the signal returns
A dispatcher dashboard with real-time adjustment suggestions
An analytical store for comparing planned against actual

The outcome

Kilometres driven fell by 12%, and planning time from four hours to eight minutes. The six dispatchers moved from manual planning to exception handling.

On-time delivery rose from 89% to 96%.

Client: National transport operator; 420 vehicles, 11 depots
Duration: 8 months
Team: 5 people

Stack

GoPython / FastAPIKotlin (Android)KubernetesClickHouse

★★★★★

The weekly report showing budget spent and remaining was, honestly, why we chose them. No other vendor offered that without being asked.

B. N. Chief Executive Officer, National transport operator

Related projects

Financial services Read the case study

A lending platform that decides in under three minutes

We replaced an approval process that took two working days with an automated flow that returns a decision in under three minutes, keeping human control where it matters.

−94% time to decision

+21% completed applications

99.97% availability in year one

Retail and e-commerce Read the case study

Unified stock for 74 stores and the online platform

Stock synchronisation between stores and the web shop moved from nightly updates to propagation in under ten seconds, eliminating cancelled orders.

−94% cancelled orders

14,000 concurrent users

40 ms median search time

Healthcare Read the case study

Medical scheduling for 22 clinics in a single system

We replaced six scheduling systems inherited through acquisitions with a single platform, cutting average phone booking time by 62%.

−62% booking duration

47% self-service bookings

−7 pp no-show rate

Got a system to build — or to rescue?

Book a 30-minute call with an architect, not a salesperson. We will tell you honestly whether we can help, how long it would take and what it would cost.

Book a call See our work