The Question That Said 'Wait' — And How I Found the Right Answer for Local Dialect
Building a language learning app for Kenyan and African languages required not just technical skill, but also deep cultural understanding. This article shares the moment when the AI assistant paused and helped find a more accurate way to structure lessons.
When I first started working on Local Dialect, I had a clear goal: to create a learning platform that focused on Kenyan and African languages beyond the dominance of English and Swahili. I wanted the app to be accessible, engaging, and rooted in real conversational contexts. But the first challenge came early — structuring the lessons in a way that felt natural for self-paced learners, not just a series of grammar drills.
I turned to Claude for help, asking how to design a lesson flow that would keep users engaged without overwhelming them. The response was helpful but vague. It suggested that the structure should be modular, but didn't offer a clear path forward. I found myself stuck — how to break down complex language concepts into manageable, interactive pieces?
After some back-and-forth, I tried a different approach. I asked Claude to imagine what a user might need when learning a new language — not just vocabulary, but the ability to communicate in real-life situations. This time, the response was more concrete. It suggested a lesson format that combined vocabulary with situational dialogues and interactive exercises, such as role-playing or matching words to images.
That idea clicked. I started building the first set of lessons with that structure in mind. I included dialogues that reflected everyday conversations — like ordering food at a market or asking for directions — and paired them with interactive exercises that required users to choose the right word or phrase. The result was more engaging than I had anticipated. Users could see how language worked in real contexts, not just as abstract rules.
What I learned from that experience was that building with AI is not just about asking for solutions — it's about asking the right questions. Sometimes, the AI doesn't give you the answer immediately, but it helps you refine the question until you find the right path.
If you're building something and feel stuck, try rephrasing your question. AI collaboration can be a powerful tool, but it needs clear, thoughtful input to guide it in the right direction. And if you're interested in seeing how this approach works in practice, you can explore the Local Dialect platform at dialect.wambugumartin.com.
Subscribe to the newsletter at https://wambugumartin.com/subscribe to get practical insights like this one, sent twice a week — every Monday and Friday.
