Abantu AI, a Kenyan startup introduces technology that can translate anything into indigenous African languages

Abantu AI, a Kenyan firm, has developed a deep learning engine for natural language processing (NLP) that can translate between major international languages and indigenous African languages.

Abantu AI was founded in September by James Mwaniki, who is also the CEO of MoVAS, a company that provides micro-lending services to telecom companies in Africa and Asia. It is developing AI-driven linguistic solutions for translating speech and text from local African languages to English and other international languages.

Currently, the deep learning model can translate from most major world languages into Kikuyu and Kiswahili, which are spoken in Kenya and East Africa. Translation into other African languages is currently underway.

“My main motivation for this project is to create a tool that will help caregivers and health workers better understand their patients in situations where the patients do not speak English or Kiswahili, or where the health workers do not speak or understand English or Kiswahili well enough,” Mwaniki said.

NLP, he said, has developed significantly elsewhere in the world, but has been lagging behind in Africa.

“Africa is a linguistically diverse continent with a significant population unable to speak in international languages such as English and French. As a result, such persons find it difficult to consume and transmit knowledge to and from the outside world without the assistance of a third party,” Mwaniki explained. “We identified a gap in the market and decided to design tools to fill it.”

Abantu AI, which is self-funded, has created a functional proof of concept and got “better than expected” market feedback from the media, government, and health care sectors.

“Our next step will be to enhance our products, adjust them to specific industry demands, and market them,” Mwaniki said.

With that in mind, the company is currently embarking on a fundraising campaign.

“We are primarily interested in grants, but we are open to alternative forms of support.” “So far, we’ve received funding from Amazon that have helped us offset the costs of training and hosting our AI models, which is often the most expensive element of the AI sector,” Mwaniki added.

By the end of the year, Abantu AI hopes to expand its services to other African countries. Clients join up for the startup’s services on a use or recurring basis, and the company makes money through subscriptions.

“We also have clients who have requested custom-made services, so this is a new income strategy for us as well.” We are not yet at the revenue realization stage because much of the work has been focused on establishing proofs of concept and testing the market, therefore our clients are still on a trial basis,” Mwaniki explained.

He explained that AI was still a new sector globally, and much more so in Africa, and that it came with its own set of challenges.

“It necessitates a set of abilities that are hard to come by locally.” It is also multidisciplinary, necessitating a broad range of knowledge. The tools for building and training models are not highly standardized, and getting any model effectively trained requires a steep learning curve. Then there’s the issue of training models, which is extremely expensive for most firms,” Mwaniki explained.

“All of these variables combine to raise the entry-level barrier for many businesses, which we hope to solve in the future by sharing our learning processes with local institutions and organizations interested in entering the field.”