Meta's and OpenAI's new AI assistants are spreading fast in the West, but expanding into Africa will hit language, infrastructure and regulatory hurdles.
Meta's Muse was downloaded 9 million times in just over a month; OpenAI's Dots starts at $100 a month. Both companies want to take these restaurant-booking, grocery-buying AI helpers to the whole world.
Speaking the local language is harder than writing code
South Africa has 11 official spoken languages. When local startup ManyHands built its assistant Manny, it found getting the AI to converse in all of them extremely difficult. English speakers would benefit first, because large language models underperform in most other languages.
Even trickier: being able to talk is not the same as knowing the rules. Across much of Africa, haggling, social ties and religion are part of every deal. Professor Arora of Utrecht University warns that if assistants follow Western norms, local users will have to adapt to the assistant instead.
The network becomes a bottleneck before the intelligence does
One task for an AI agent requires constant back-and-forth between servers, websites and the user. Cisco measured up to 450% more traffic per task than a human; over the past year, daily AI-agent requests on Cloudflare's network grew more than 1,700%.
Places with spare capacity, like parts of East Africa, will cope for now. But Niger, Burkina Faso and eastern Congo already have poor connectivity, and road construction regularly cuts fibre cables. If a submarine cable fails, the cleverest assistant cannot reach users to get the next approval.
Uneven foundations mean assistants widen the gap
Only 36% of Africans used the internet in 2025, against 92% of Europeans. City dwellers in Africa were 2.6 times more likely to be online than rural residents. Carnegie fellow Jane Munga says agents could deepen inequality, since most people cannot afford paid services like Dots.
For a local business to let an AI assistant take orders, it must move online, manage inventory, and set up digital payments — all costing money and skills a market vendor does not have.
Why it matters
AI assistants work in the West because payments, logistics and networks quietly hold everything up. In emerging markets, one of those layers is often missing. Tech companies that only translate the interface while ignoring the foundations may end up with agents that arrive and find no one able to catch them.
Curated from high-quality sources, with concise summaries and key takeaways.
