Adapting AI to Africa’s classrooms means designing tools that are offline-capable, low-bandwidth friendly, capable of switching between local languages, embedding local culture, and built with local educators’ input. Teachers may not have been trained to use AI tools or to understand their limitations—about bias, about privacy, about ethical use. Only then can localised AI tools run reliably. Localised AI in education isn’t just about better test scores (though it can help with that) — it’s about dignity, relevance, and equity. As more localised AI models emerge, as policies take shape, and as infrastructure improves, classrooms across Africa can shift from “chalk-and-talk” to interactive, adaptive, culturally resonant learning environments.