The Day the App Said 'Not Ready' — And How I Found the Right Question for Mwalimu Cosmetics

Every product I build starts with a problem. For Mwalimu Cosmetics, the problem was clear: an e-commerce platform that could support a local cosmetics brand in Kenya. I wanted to create a seamless experience for customers and a manageable tool for the store owner. But there was one moment—early in the process—that made me question everything.

I was working on the payment integration when the app said, 'Not ready.' It wasn't a crash or an error message. It was a quiet refusal to process the transaction, like the system was waiting for something I hadn't provided. I had followed the AI's guidance on structuring the payment flow, and yet, the system didn't respond as expected. I was frustrated, but I wasn't ready to give up.

That moment taught me something important: the AI is a tool, not a replacement for understanding the real user. I had assumed that following the AI's logic would be enough. But I hadn't considered the nuances of the Kenyan payment landscape—specifically, the need for multiple payment gateways and the way different banks handled transactions. I had to go back, not just to the code, but to the people who would be using the platform.

So I reached out to the store owner, asked her about her experience with online payments, and listened. She mentioned that customers often used M-Pesa, but there were also times when bank transfers were necessary. The AI had suggested a single integration, but the real world required flexibility.

This led to a change in the project. Instead of pushing forward with the AI's first recommendation, I revisited the architecture and added support for multiple payment methods. It took time, but it made the platform more robust and user-friendly. The result was a system that didn't just work—it worked for the people who needed it most.

Building AI-powered tools is about more than writing code. It's about listening, learning, and sometimes, admitting that the AI's first suggestion might not be the right one. That moment with Mwalimu Cosmetics was a turning point in how I approach every project now.

If you've ever felt like your AI tool didn't quite understand your needs, you're not alone. I've been there too. But every time I hit a wall, I've found that the solution lies not in the code, but in the people who use it. What's one assumption you've made about your project that might not be right? What would happen if you asked the people who need it most?

Subscribe to the newsletter for more insights like this, delivered every Monday and Friday. https://wambugumartin.com/subscribe