The Chat That Said 'Try Again' — And How I Found the Right Prompt for Local Dialect
Building a language learning app for Kenyan and African languages presented unique challenges, especially when it came to designing interactive lessons. A pivotal moment came when AI collaboration suggested a new approach that transformed the learning experience.
Creating Local Dialect was about more than just building an app — it was about making sure the lessons felt real, not just academic. I had a rough idea of how the lessons should flow, but when I first tried to outline them with AI, the results felt flat. The chatbot suggested something I hadn't considered: grounding the lessons in real-life conversations, not just grammar drills. That idea stuck, and it changed the entire approach to content design.
I wanted the lessons to feel like a natural extension of how people actually use these languages — in markets, in homes, in everyday exchanges. But translating that into structured content was tricky. I tried several prompts, but each time the AI-generated lessons lacked the authenticity I was looking for. Then, one evening, I tried a different approach. Instead of asking for structured lessons, I asked the AI to 'write a short conversation between two people in Nairobi discussing a local event, using Swahili and a regional dialect.' The response was immediate and rich with cultural nuance. It was exactly what I needed.
From that moment, I shifted the focus of the AI collaboration. Instead of asking for abstract lesson structures, I asked for real-life scenarios. The AI responded with a series of short, interactive dialogues that mirrored the way people actually speak. These became the building blocks for the lessons. I used them as templates, then refined them with human input to ensure they were both accurate and engaging.
The result was a learning platform that felt alive — not just a list of vocabulary and grammar rules, but a space where learners could practice speaking in a way that mirrored their own lives. That was the moment the app felt like something real, not just a prototype.
Building with AI is not always about getting the right answer the first time. It’s about listening, adjusting, and finding the right prompt that unlocks the kind of creativity and insight you need. For Local Dialect, that moment came when the AI suggested a different way of thinking about language learning — and it made all the difference.
If you're building something and the first attempt doesn't feel right, try a different prompt. Sometimes, the answer you need is just waiting for the right question.
