Kathy Gibson reports – The true value of artificial intelligence (AI) in Africa could lie in the ability to develop local language models that allow people across the continent to access services in their mother tongue.
Ben Cheick Haidara, deputy-CEO and chief operating officer of Orange Middle East and Africa tells delegates to the Africa Tech Festival in Cape Town that Africa should be looking to practical uses of AI technology.
Globally, AI is set to have a net impact of about 1,2% of total global GDP – so the technology is not hype, he asserts.
“It means that AI could add value to businesses, states, and citizens.”
Some of the business use cases are already unfolding in the form of personalised customer relationship management (CRM), smart bundling, contextual upselling and more. “These use cases can impact revenues while delighting customers with proactive and predictive customer care,” Haidara says.
For governments, AI could help them to be more efficient, with more optimisation in public services. “For example, a chatbot could provide 24/7 services which would make a big difference for citizens.”
But the biggest impact that Africa could see, he believes, is the ability for citizens – or customers – to interact with models, moving from a simple request for information to a digital conversation.
Orange uses AI in its operations for predictive maintenance, dynamic resource allocation, and smart investment. Employees also use generative AI (GenAI) internally to boost productivity. And AI is useful for personalising customer offers.
One of the use cases that Orange has developed is the ability to build local language models. “We are working to understand real customer conversations in local languages,” Haidara explains.
Another use case the telco is offering customers is a chatbot dedicated to the upcoming African Cup of Nations, with an avatar that can answer users’ questions.
“We believe in AI,” Haidara says. “We think it’s the way to go, and we are supporting initiatives to grow it and make it real.”
There are questions about whether AI can scale in Africa, and Haidara points out that the bottleneck is no longer connectivity. “The coverage issue no longer exists. The GSMA says the coverage gap is now less than 10%.
“Today the real issue is penetration and the usage gap, so we should work a bit more on making devices affordable.
“We could increase inclusion with local language models as well. There are more than 2 000 languages in Africa and we feel being able to speak to people in their own language will be the most powerful inclusion engine.
“We can truly reach those people who are left on the ground.”