When the Code Spoke — And What It Revealed About the People Behind It
Building AI tools for everyday life in Kenya has been a journey of discovery, frustration, and pride. Each project has taught me something new about the people who use them — and about myself as a builder.
"When the Code Spoke — And What It Revealed About the People Behind It Building AI tools for everyday life in Kenya has been a journey of discovery, frustration, and pride. Each project has taught me something new about the people who use them — and about myself as a builder. There was a moment during the development of Local Dialect when the code worked — not just in theory, but in practice. I was testing the lesson flow for a regional Kenyan dialect, and the AI-generated content actually made sense. It wasn’t just a string of words; it felt like a real conversation. That moment taught me that AI, when used correctly, can help bridge the gap between technology and the human experience. But it also reminded me that I had to be careful — not all AI outputs are perfect, and sometimes the human touch is still needed to make sure the content is culturally and linguistically accurate.
Building Mwalimu Cosmetics was a different kind of challenge. I had to make sure that the e-commerce platform wasn’t just functional, but also intuitive for users who might not be familiar with online shopping. AI was helpful in designing the initial structure of the platform, but it was the real-world feedback from users that shaped the final product. One of the most surprising things was how much the platform’s success depended on the simplicity of the user interface. People wanted to buy, not to learn how to buy — and that was a lesson I didn’t expect to learn from a machine.
With Arise & Shine Transporters, the hardest part was figuring out how to make the AI-powered anomaly detection system work in real-time. At first, the system flagged too many false positives — it was too sensitive. I had to go back and adjust the parameters, but more importantly, I had to understand the context in which the system would be used. It wasn’t just about detecting route deviations; it was about understanding the drivers’ behavior and the business’s needs. That process taught me that AI is only as good as the data it’s trained on — and that data has to be deeply rooted in the real-world problems it’s meant to solve.
Rev & Learn was another project that surprised me in unexpected ways. I had assumed that the age-grouped question packs would be the most difficult part of the development, but it turned out that making the content engaging and age-appropriate was the real challenge. AI helped structure the questions, but it was the human input that made the difference. I realized that building for children wasn’t just about making things fun — it was about making them meaningful, and that required more than just clever algorithms.
Planner, the AI-powered planning assistant, was the first product where I really felt the power of AI in a different way. It wasn’t just about generating a to-do list — it was about helping users think through their goals and figure out what steps to take next. The AI didn’t just suggest tasks; it helped users prioritize and understand dependencies. That was a moment of pride, because it showed that AI could be a real thinking partner, not just a tool.
Each of these projects has left its mark on me — not just as a builder, but as a person who is constantly learning and adapting. There were times when I felt overwhelmed, and times when I was proud of what I had accomplished. But through it all, I’ve come to
