The Wrong Map — And How I Found the Right Route for Arise & Shine
When building the Arise & Shine Transporters platform, I assumed that GPS data would always align with local road networks. But one day, the map said something different — and it changed everything.
I assumed that GPS data would always match the roads in Kenya. That was my mistake. For weeks, I watched the fleet tracking system show trucks moving along routes that didn’t exist — not because the GPS was broken, but because the map data didn’t account for local roads, detours, or seasonal road closures. I had relied on a generic mapping API, thinking it would handle everything. It didn’t. It led to confusion, wrong pricing, and frustrated users who saw trucks taking longer routes than expected.
The moment of realization came when a truck driver called me. He said, 'Sir, the app says I’m on Mombasa Road, but I’m on a dirt path near Thika. Why is the map wrong?' That call stopped me in my tracks. I had built a tool that claimed to know where trucks were, but it was built on assumptions that didn’t match the reality of Kenyan roads. It was a humbling moment — and one that forced me to rethink the whole approach to mapping.
What changed? I switched to using OpenStreetMap Nominatim for geocoding, which relies on crowdsourced data that better reflects local roads and conditions. I also added a feature to allow drivers to report road issues in real time. This gave the system a way to learn and adapt, making the app more accurate and trustworthy.
This mistake taught me that AI tools, no matter how powerful, can’t replace the nuance of local knowledge. When I built the AI-powered ops copilot, I made sure it could reference real-time road conditions and driver feedback. That made all the difference — and it made the app feel more like a partner than a tool.
I’ve learned that mistakes like these are not failures. They’re moments of clarity. They show us where our assumptions are wrong and where we need to listen more closely — to drivers, to users, and to the data that reflects the real world.
If you’ve ever built something and found out that your assumptions were wrong, you’re not alone. It happens to all of us. The key is to keep learning, keep adapting, and keep building better tools — even when the map says 'no.'
Subscribe to our newsletter for more insights like this — sent every Monday and Friday. [https://wambugumartin.com/subscribe]
