AI & the future of telecoms and connectivity

A mobile network with tens of thousands of transmitters now handles more and more of its own management. AI can see from the traffic that a stadium will fill up within the hour and has extra capacity ready. When a component falters, the system reroutes the traffic and restarts it. Many operators are working towards autonomy level 4 as defined by the industry body TM Forum, at which the network manages itself within defined situations. Expansion is planned with AI as well: it combines maps, traffic data and coverage complaints, and shows which new mast or fibre route reaches the most households per euro invested. The next step is AI agents that are given a goal rather than rules, such as keeping coverage in a neighbourhood up to standard with as little power as possible.

A close-up inspection used to mean a technician climbing the mast. Now a drone circles it in a few minutes and takes hundreds of photos. Image recognition marks visible defects such as rust, misaligned antennas or a damaged cable, and the images are turned into a three-dimensional model that shows an engineer whether there is room for a new antenna. Maintenance shifts from a fixed schedule to the places where something is actually wrong, and there is less climbing at height. The next step: drones that take off by themselves from a docking station, with no pilot on site, and a digital twin of every mast that keeps itself up to date.

A mobile mast uses a lot of power, even when hardly anyone is calling or streaming. AI forecasts how much traffic each transmitter will carry over the coming hours. When an office district goes quiet at night, the system temporarily switches off some antennas or frequencies and neighbouring transmitters pick up the traffic. When demand rises, they are back on within a short time. The cooling in the cabinets next to the mast is managed more efficiently too. The next step goes further. For 6G, near-zero power use by sleeping radio components is a design goal. Networks will also match their electricity use to sun and wind, partly with batteries at the masts. Telecoms then becomes a flexible customer that helps the grid.

Ports, factories, mines and airports are getting their own mobile networks, with their own access rules and their own agreements on capacity. Among steel containers and machinery, such a network is often more reliable than wifi. Driverless vehicles, remotely operated cranes, cameras and thousands of sensors stay connected, and AI adjusts the network when a robot needs more bandwidth. In a port, container carriers drive between quay and storage without a driver, and a crane operator works the crane from an office. The next step: factories where wireless links replace the cables between machines. A production line can then be switched to a different product overnight.

When the network handles much of its own management, the work of the people who run it changes. The engineer who used to adjust settings transmitter by transmitter now writes the rules and software that tune thousands of transmitters at once. New services bring roles that barely exist yet. The network agent designer decides which goals and permissions an AI agent gets. The steward of network sensing decides who may see what once the network becomes a sensor. The quantum network engineer builds links that exchange keys using particles of light, and the holographic service designer makes three-dimensional calls possible. Experience keeps its value: someone who has spent years fixing outages knows the exceptions a model still misses.

A network operations centre receives thousands of alarms every day. In a team of people and AI, the system filters those alarms, merges reports with the same cause into one incident and points to the likely source. It fixes a simple fault itself; for a bigger problem, an assistant proposes a solution and the engineer approves it or steps in. AI takes on the repetitive and the deeply analytical work. The engineer does what people are good at: judging, choosing and leading the recovery after a major outage. The picture for 2040: a network centre with nobody in it at night, and engineers who spend the day on the anomalies the system could not explain by itself.

A network that runs itself needs leaders who are out in front. They take the initiative on AI and new technology and show where the company is heading: which work the network will soon do by itself, and where engineers, field technicians and customer service make the difference. That picture of the future has to hold up in the network centre and at the foot of the mast. Leaders make it tangible with a first application everyone can see working, and give teams room to experiment. That is how they get their people building tomorrow's networks together.