The Moment the Map Said 'No' — And How I Found the Real Problem

It was a late afternoon in Thika, Kenya, and I was standing next to a truck that had just returned from a delivery. The driver, a man with decades of experience on the road, pointed at the GPS screen on his phone and said, 'This isn't right.' The map showed a straight line from point A to point B, but the driver knew better. He had taken a different route — one that wasn't captured in the system. The problem was clear: the digital map didn't match the physical road network. And that discrepancy was costing the business real money.

The realization hit me like a truck. The platform I was building for the sand and aggregates supply business was based on the assumption that digital maps were accurate. But in reality, the roads in East Africa — especially those outside major cities — were often unmarked, poorly maintained, or simply not represented on standard mapping tools. This meant that the system couldn't accurately calculate delivery costs, track truck movements, or detect route deviations. It was a critical flaw that could undermine the entire project.

I sat down with the business owner and the driver, asking questions. What routes do you usually take? How do you know when a truck is off course? What happens when fuel receipts don't match consumption? The answers painted a picture of a system that relied on human judgment and experience — not digital tools. The GPS was a helpful addition, but it wasn't the solution. The real solution had to account for the nuances of local geography and the realities of manual logistics.

That's when I turned to AI. I used tools like Codex and Claude to explore alternative approaches to mapping and route detection. The AI didn't just help me find a better way to represent the roads — it helped me rethink the entire approach to fleet management. Instead of relying on static maps, the system now uses real-time GPS data combined with historical route data. This allows the platform to adapt to the actual conditions on the road, not just what's on the map.

The moment I realized the problem was the same moment I began to see the solution. It wasn't just about building a better app — it was about understanding the people who used it. The drivers, the business owners, and the logistics teams all had knowledge that the digital tools weren't capturing. By integrating AI into the development process, I was able to build a system that respected that knowledge and built on it, rather than replacing it.

Now, the Arise & Shine Transporters platform doesn't just track trucks — it learns from them. It adapts to the roads they take, the routes they know, and the challenges they face. The map no longer says 'No' — it listens, and it learns.