Africa-focused voice technology company Intron has launched Sahara v2.5. This is an updated voice AI platform designed to better understand how Africans naturally communicate, including conversations that switch between languages.
The release introduces bilingual speech recognition or code-switching, across 12 African languages, including Swahili, Zulu, Hausa and Luganda. Intron says its benchmarks show Sahara outperforming models from Gemini, ElevenLabs and Meta across the 12 languages tested.
The company has also introduced what it describes as the world’s first African trilingual speech-recognition model supporting switching between Kinyarwanda, English and French. Intron says the technology is based on proprietary algorithms for which it has filed US patents.
Sahara expands African language coverage
Beyond code-switching, Sahara v2.5 adds speech recognition for Nupe, Kanuri, Nigerian Fulfulde, Tigrinya, Kikuyu, Dholuo and Somali. This brings the platform’s total African-language coverage to 31 languages.
The platform also supports language-mixing for text-to-speech and voice agents across 13 languages. This allows businesses to build voice bots and assistants that can respond more naturally to users.
Intron says Sahara recorded an average word error rate (WER) of 34.3% across 12 African languages, compared with 53.8% for Gemini 3.6 in its published benchmarking. A lower WER indicates fewer transcription errors.
The company says the improvements could have applications across contact centres, healthcare, financial services, government and legal services, particularly where conversations naturally move between English, French and local languages.
Sahara is already being used by more than 40 enterprise customers across Nigeria, Kenya, South Africa, Uganda, Rwanda and Ghana.
Among existing deployments, Intron says Branch International’s Sahara-powered collections agents recovered more than ₦1.2 million in delinquent loans in one week.
In healthcare, Intron says a 14-minute Swahili-English doctor-patient consultation in Nairobi can be converted into a structured clinical note in less than 30 seconds.
“Code-switching was one of the biggest problems that consistently came up for clients deploying real-world voice AI,” said Tobi Olatunji, CEO of Intron. “Africa needs AI built for how Africans really speak. People should not have to translate themselves for a machine, flatten their accent, avoid local expressions or repeat only the English part of what they said.”
For these and more stories, follow us on X (Formerly Twitter), Facebook, LinkedIn and Telegram. You can also send us tips or reach out at info@techarena.co.ke.
Also Read: How Intron’s African Voice AI Is Transforming Courts, Clinics, and Call Centres