He Built an AI that Finally Understands the Many Accents of African Doctors

For decades, doctors across Africa have been drowning in a sea of paperwork. In many hospitals, patient records are still handwritten, or doctors spend hours typing out notes after a long shift. While voice-to-text technology has existed for years, it has a glaring "accent problem." Standard AI models, trained primarily on American or British voices, often fail to understand a doctor from Lagos, Nairobi, or Accra. This digital bias was a source of deep frustration for Tobi Olatunji, a Nigerian medical doctor who transitioned into a machine learning engineer. He saw that the inefficiency of documentation wasn't just an administrative headache; it was a barrier to quality patient care.

Tobi’s journey with Intron Health began with the realisation that for AI to be useful in Africa, it had to be built for Africa. He didn't want to just "tweak" an existing Google or Amazon model; he wanted to build a foundational tool that respected the linguistic diversity of the continent. Starting in 2020, Tobi and his team began the painstaking process of collecting thousands of hours of clinical speech data from across Nigeria and beyond. They recorded doctors with various accents, Yoruba, Igbo, Hausa, and more, explaining complex medical conditions. This data became the "training ground" for an AI that can now recognise over 200 African accents with nearly 95% accuracy.

Building a "Maker" business in the AI space requires more than just code; it requires a deep understanding of the environment. Tobi designed Intron Health to work even with the background noise of a busy, non-air-conditioned clinic and to function seamlessly with existing Electronic Health Record (EHR) systems. In an interview with TechCrunch in 2022, Tobi explained that by reducing documentation time by up to 70%, he is effectively "giving doctors back their time" to spend with patients. As of 2026, Intron Health has been adopted by over 30 hospitals across West Africa, and Tobi is now expanding the model to include French-speaking and Swahili-speaking medical professionals.

Tobi Olatunji represents a new wave of African innovators who are not just users of Global North technology, but creators of culturally specific AI. His work proves that the "future of work" in Africa isn't just about automation; it's about making technology more human and more inclusive. For Tobi, the goal is simple: a world where an African doctor can simply speak their diagnosis, and the technology listens—no matter how they sound.

Lessons for Budding Makers

Tobi Olatunji’s success with Intron Health provides two powerful lessons:

1. Identify the "Blind Spot" of Global Tech: Global tech giants often build for the majority, leaving massive "edge cases" (like African accents) underserved. Your greatest opportunity as a maker is to find where global tech fails in your local context and build the missing bridge.
2. Data Sovereignty is Key: In the age of AI, the person with the most relevant data wins. Tobi didn't just write code; he built a unique dataset of African medical speech that no one else had. If you want to build AI, start by collecting the specific, local data that the rest of the world has ignored.