MTN turning its fibre network into a sensor network
MTN is deploying technology that automatically detects when physical disruptions affect its fibre Internet network, which reduces the time it takes to dispatch technicians to fix problems.
This was revealed by group chief technology officer Charles Molapisi, who detailed how the telecom giant expects to create R30 billion in value through AI adoption and products.
“We lose a lot of service today because of fibre cuts. We have a lot of challenges in terms of excavation, road maintenance, and all of that,” he said during MTN’s Capital Markets Day.
“We are now turning the fibre network into a sensor network, which means it actually observes the micro-vibrations on the network and picks up patterns.”
He explained that they employed agentic AI, which began to learn how the fibre network is supposed to behave, and alert technicians to fluctuations in connectivity.
Molapisi said the technology would shorten the time required to diagnose and repair damage to the fibre network, which previously took about 8 hours.
“When people are mobilising, and before they put that TLB down to dig, we are already there to manage and support ourselves,” he said.
Mazen Mroué, CEO of digital infrastructure at MTN, revealed that the group expected its fibre footprint across Africa to expand significantly in the coming years.
The telecom’s fibre network will expand to between 420,000km and 560,000km by 2030 as it targets 100 million new data customers in Africa.
“When we look at the fibre core, our fibre footprint is expected to grow, increasing by three times from our current 140,000 kilometres over the next five years,” said Mroué.
“This will be backed by subsea capacity, which is important for connecting the African continent to the rest of the world.”
He said that the group expected its subsea cable capacity to double heading into 2030. The group manages subsea and terrestrial wholesale capacity through its subsidiary, Bayobab.
R30 billion in value from AI
Molapisi was outlining MTN’s group-wide AI strategy, which included internal adoption across the company to increase efficiency and the release of new AI products.
These new products will be geared to both regular customers and businesses, and will include the creation of small language models and the leasing of compute from its data centres.
“The promise we are making, and this is our ambition, is that in three to five years, through AI, we must deliver R30 billion in value creation,” said Molapisi.
He said that R15 billion of that amount would be created by AI Inside, the company’s plan to adopt AI across its internal processes. This adoption was a near certainty for MTN.
Meanwhile, the rollout of AI products to customers and businesses accounted for the remaining R15 billion, which was a riskier goal, he said.
“Depending on what happens on B2C, who knows what the number will become, but we are a little bit more certain in terms of our ability to execute on AI Inside.”
Molapisi said the company was deploying smart agents across its network to help its technicians with incident management.
These agents would also be used to tune some parameters through an approval process on the network, with Molapisi calling it “a significant gain for MTN.”
“This is what every telco in the world is doing, and the assurance I am leaving with the investors today is that we are not far behind; I think we are moving at a faster rate.”
Agentic AI for SIM registration in Nigeria
A significant portion of the company’s AI innovations for customers is focused on Nigeria, its largest market by revenue.
In the West African country, the group has rolled out agent-aided SIM registrations, where 13 AI agents are helping customers register their SIM cards.
Previously, the work was handled by around 200 employees across the country who handled registrations.
“13 automated agents work together to distribute the load. When I say agent, I mean agentic AI. We expect this solution will scale across many markets,” he said.
“We expect it should give us proper upside in terms of improvement, accuracy, and compliance.”
Molapisi said that human employees would be kept in the loop to address governance and control issues related to registrations.
Loading ...