Nigerian voice technology startup Intron is reengineering the future of voice AI by focusing on a long-ignored challenge in global tech: understanding Africa.
With a proprietary system trained on over 30,000 hours of voice data from more than 30 countries, Intron is now expanding its reach far beyond healthcare, bringing its innovation to finance, telecoms, legal, and government services.
Founded in 2020 by Tobi Olatunji, a radiologist-turned-technologist, and Olakunle Asekun, Intron first made headlines for developing Africa’s first clinical speech recognition platform.
The initial system demonstrated a 92% accuracy rate on medical speech recognition, a breakthrough that outperformed many international models, especially when dealing with African accents.
Following a $1.6 million pre-seed funding round in July 2024, backed by investors such as Ventures Platform and Future Africa, the startup has rapidly broadened its capabilities. Today, it offers real-time voice-to-text services across multiple industries through its flagship voice AI platform, Sahara.
Sahara is built on a unique dataset comprising over 3.5 million voice clips from more than 18,000 speakers across 30+ African countries. The dataset powers Intron’s patented AccentMix algorithm, which enables the system to process over 300 African accents and dialects.
It also demonstrates high performance with North African and Arabic-accented English, outperforming several frontier AI models.
This depth of linguistic exposure has positioned Intron to build two groundbreaking AI models. The forthcoming Sahara-Titan will be capable of understanding, transcribing, and translating between 20 of Africa’s most widely spoken languages, including Swahili, Hausa, and Zulu. Meanwhile, Sahara-Primus will generate fluent, natural-sounding audio in these languages, paving the way for more inclusive and localized digital services.
“Most voice AI systems today are trained on Western data. That leaves billions of users in Africa poorly served,” said Olatunji in an official company statement. “Instead of complaining about model bias, we’re choosing to build better models that reflect our voices, our languages, and our realities.”
Why It Matters
Intron’s work arrives at a critical time when Africa’s digital economy is rapidly expanding. According to a report by the International Finance Corporation (IFC), Africa’s internet economy could contribute nearly $180 billion to GDP by 2025.
Voice-based applications, from mobile banking and e-commerce to public service access, are expected to play a central role, particularly in regions where literacy or language barriers persist.
By centering African linguistic identity within cutting-edge AI development, Intron is not only addressing historical bias but also unlocking new commercial and societal potential. Its innovations suggest a future in which African voices are no longer excluded or mistranslated by technology but are instead driving it forward.
Talking Points
Africa is tired of being an afterthought in global tech—and it’s showing. Intron’s voice AI revolution is a bold statement: Africa is no longer waiting to be “included” by Silicon Valley. We are building our own. This is a wake-up call to global developers: your models don’t understand us, and we’re no longer asking politely.
Big tech’s blind spot has been Africa’s opportunity. For years, global voice assistants like Siri, Alexa, and Google Assistant have stumbled over African names and accents. The bias wasn’t just technical; it was cultural. Intron proves that when Africa builds for Africa, innovation doesn’t just catch up; it leaps. Why train your model on 10 countries when you can build it on 30+ African nations?
This is bigger than tech—it’s a cultural power move. Language is identity. When AI fails to recognise your voice or mispronounces your name, it’s not just a glitch; it’s erasure. Intron is reclaiming space not just in technology, but in global identity. The West shouldn’t just be impressed; they should be worried. Africa has entered the chat.
