Transport for London is upgrading one of Europe's largest traffic signal networks with artificial intelligence, aiming to reduce delays by up to 14 percent. The technology is quiet, almost invisible, and that is exactly what makes it worth watching.
Earlier this year, London received a ranking no city wants. It was named the slowest capital city for traffic in the world. Days later, Transport for London unveiled its answer: a five-year strategy called London on the move, the city's first ever network-wide traffic management plan.
At first glance, it reads like a technical document about roadworks and bus lanes. In reality, it describes something more interesting. London is in the process of teaching its road network to think.
The Problem Is Bigger Than a Red Light
Congestion is often discussed as an inconvenience. For a city like London, it is an economic issue. The Mayor of London, Sadiq Khan, put it plainly when launching the strategy: congestion is annoying for everyone, and it costs the economy millions of pounds every year.
Every minute a delivery van sits in traffic is a cost to a business. Every slow bus journey makes public transport a little less attractive. Every blocked junction ripples outward through supply chains, appointments and working hours. A city's traffic system is, in a very real sense, part of its economic infrastructure.
London's version of that infrastructure is enormous. TfL manages around 6,400 automated traffic signal junctions and pedestrian crossings, one of the largest networks of its kind in Europe. Until now, those signals have largely followed programmed patterns. The new strategy changes what they are capable of.
What TfL Is Actually Changing
At the centre of the plan is an upgrade to FUSION, London's traffic control system. With more data and more computing power, the system is being developed to react faster and more intelligently to conditions on the road, adjusting signals to keep traffic flowing rather than following fixed timings. TfL estimates that these changes could reduce delays by up to 14 percent.
Alongside it, TfL is expanding its network of Vivacity cameras, which use artificial intelligence to understand how people are actually moving through the city. The cameras can distinguish between pedestrians, cyclists, wheelchair users, taxis and heavy goods vehicles, giving planners a far more detailed picture of each street. TfL is also finalising a data-sharing arrangement that would give London's boroughs access to insights from more than 1,000 of these cameras.
Buses are a priority of their own. Technology that gives buses green lights more frequently is already installed at 2,080 traffic signals, and the goal is to extend it to all 3,500 signals on bus routes by 2030. The strategy also expands the Lane Rental programme, which charges contractors for occupying busy roads during peak times, and introduces the first pan-city plan for coordinating roadworks.
None of this will appear on a skyline. Most Londoners will never see any of it directly. They may simply notice, over time, that the lights seem to know what they are doing.
Why This Matters Beyond London's Roads
There is a broader story inside this strategy, and it is about how artificial intelligence actually arrives in daily life.
Public conversation about AI tends to focus on chatbots, images and headlines. But some of the most consequential uses of the technology look nothing like that. They look like a traffic signal holding green for three extra seconds because the system knows a bus is approaching. They look like a junction adjusting itself at 7am because the data shows this Tuesday is behaving differently from last Tuesday.
This is AI as infrastructure. It does not announce itself. It simply makes an existing system work better, and at the scale of a city, small improvements multiply. A few percent less delay, repeated across millions of daily journeys, adds up to real time, real money and real quality of life.
Other cities will be watching. If London, with some of the most complex traffic conditions in the world, can make measurable gains through smarter signals, the approach becomes a template.
What Businesses Can Take From This
There is also a quieter lesson here that applies well beyond transport.
London's congestion answer is not, primarily, new roads. It is making better use of the network the city already has, by measuring it properly and letting the data drive decisions. That principle translates to almost any organisation. Before investing in something new, it is worth asking whether the existing system is actually being used well, and whether anyone is measuring it closely enough to know.
The sequence TfL is following is worth noting: first better data, then better decisions, then better flow. Most improvements, in cities and in companies, happen in that order.
The Smartest Technology Is the Kind You Never Notice
The London on the move strategy will take years to deliver, and its success will be measured junction by junction, month by month. Not every target will be hit, and congestion will never disappear from a city of London's size.
But something meaningful is happening at those 6,400 junctions. A piece of infrastructure that has worked the same way for generations is learning to respond to the city around it.
For decades, Londoners have waited at traffic lights. Soon, in a small but genuine way, the traffic lights will be paying attention to them.
