How AI is optimizing last-mile delivery in Mbeya
Mbeya's hilly terrain and rapid urban growth make last-mile delivery a unique challenge. Pacelio's AI-powered routing engine is turning that challenge into a competitive advantage.
The problem with traditional routing
Most courier platforms rely on shortest-path algorithms that work well on flat, grid-based cities. Mbeya's topology — steep slopes, narrow feeders, and seasonal road closures — demands a more intelligent approach. Before Pacelio, couriers in the Southern Highlands wasted an average of 25 minutes per trip navigating around impassable roads or backtracking from dead ends.
How Pacelio's algorithm works
Our routing engine combines three data sources in real time:
- Historical traffic patterns — learned from thousands of completed deliveries across Mbeya's wards
- Live road conditions — crowdsourced from active courier GPS feeds every 30 seconds
- Terrain weight maps — elevation and surface-type data that penalizes routes with steep gradients or unpaved sections
The result is a dynamic route that adapts throughout the day. When heavy rain makes one road impassable, the system reroutes the entire batch in under two seconds.
Measurable impact
In the three months since the algorithm went live, Pacelio has recorded:
- 30% reduction in average delivery time
- 22% more deliveries per courier per shift
- 18% decrease in fuel costs for the fleet
- 95% on-time delivery rate within Mbeya city centre
What's next
We are currently training the model on data from Dar es Salaam and Arusha ahead of our expansion. The goal is a self-improving logistics brain that gets smarter with every parcel delivered across East Africa.