TL;DR
Cambridgeshire County Council has started a £100,000 contract for AI-powered software that automatically detects potholes and other road defects using cameras fitted to inspection vehicles. The council maintains 4,600km of roads and faces a maintenance backlog of £800m.
How the system works
Cameras mounted on inspection vehicles capture footage of road surfaces, and AI identifies defects such as potholes automatically. Officers can then review the stored footage for post-inspection checks. The council says the next step is integrating the technology with its works ordering system, which would streamline the process from detection through to repair scheduling.
The contract began earlier this year, and the council reported that weekly defect numbers dropped from a peak of 2,892 in February to 830 in the week ending 29 March — though the seasonal improvement likely contributed alongside any technological gains.
The scale of the problem
Cambridgeshire’s road maintenance challenge is substantial. Council papers submitted in January put the total backlog at £800m, with £500m of that attributed to so-called soil-affected roads built on peat that buckle and cost up to four times more to repair than standard surfaces. The council leader has previously admitted restoring all roads to good condition would be “pretty much impossible.”
The government recently rated Cambridgeshire’s road maintenance as “amber.”
What this means for UK local government
Cambridgeshire is not alone in turning to AI for infrastructure management. Several UK councils have trialled similar computer vision systems, reflecting a broader pattern of local authorities seeking technology solutions for problems that outstrip their budgets. At £100,000, the contract is modest relative to the backlog it aims to help manage. The real test will be whether automated detection translates into faster, better-prioritised repairs rather than simply a longer, better-documented list of problems.
Looking forward
If the integration with the works ordering system succeeds, the approach could offer a template for other councils facing similar backlogs. The gap between identifying defects and actually fixing them remains the harder problem to solve.