AI
How UK Manufacturers Are Quietly Adopting AI on the Factory Floor
Predictive maintenance adoption among UK manufacturers more than doubled in a year. Here's what the data shows — and why skills, not budget, is now the bottleneck.
By Mark Foster · Contributor & Tech Writer · Published
Last updated
Mark Foster contributes to British Business Review as an independent journalist and is not a member of staff.
The number worth sitting with: predictive maintenance adoption among UK manufacturers more than doubled in a single year, from 9% to 22%, according to a survey of over 600 senior manufacturing and maintenance decision-makers across the UK, US and Germany, commissioned by test-and-measurement firm Fluke and conducted by Censuswide, published in May 2026. Reactive maintenance — fixing things after they break — fell from 42% to 26% over the same period among UK respondents. That's a genuine shift in how factories operate, and it's happened almost entirely without the fanfare that's followed generative AI everywhere else.
Why predictive maintenance is the easy case
Of all the industrial AI use cases, predictive maintenance has the cleanest business logic, which likely explains why it's the one that's actually scaling. Sensors on equipment — measuring vibration, temperature, sound, power draw — feed a model trained on historical failure patterns, and the model flags early signs of the kind of wear that has previously preceded a breakdown. Unplanned downtime on a production line is expensive in a way that's easy to underestimate until it happens: lost output, idle labour, and rush-order penalties stack up fast, and even a modest reduction in surprise stoppages tends to pay back the sensor and software investment quickly.
The Fluke survey also points to where the money is actually going now versus where it went a year ago. Investment in "exploratory AI" fell from 53% of respondents in 2024 to a smaller share this year, replaced by spending on Generative AI (38%), cybersecurity (37%), Industrial AI (34%) and data management (32%) — in other words, a shift from pilot projects toward tools tied to day-to-day production. Nearly three-quarters of surveyed organisations now put 16–30% of their maintenance budgets toward new technology.
The constraint isn't money anymore — it's people
Here's the part that undercuts the tidy "AI adoption is accelerating" narrative: the same survey found that skills shortages, not funding, are now the primary obstacle to further adoption, accounting for roughly 77% of all reported barriers — split across general knowledge shortages, workforce skills gaps, lack of in-house expertise, and skilled labour scarcity. Fluke's chief product officer Vineet Thuvara put it plainly: predictive maintenance is "no longer a future ambition," and the harder problem now is scaling it across an organisation rather than running it in isolated pilots.
Free newsletter
The British Business Brief
One email a week on UK business, regulation and capital: what moved, who paid for it, and what it means. No hype.
Where the sector leads and lags
Adoption isn't even. Automotive and electronics manufacturing lead the field, with around 60% of mid-to-large automotive firms using AI in some production capacity, according to a separate benchmarking study by IDS-INDATA — the highest of any sector it tracked. Jaguar Land Rover has been cited as a specific example, running AI-powered production analytics across 128 sites to catch anomalies in real time. Aerospace, defence and pharmaceuticals follow at a slightly lower level of maturity, while food and beverage manufacturing, per the same study, has the most ground still to make up.
Why this hasn't made headlines
Part of the reason industrial AI gets so little press relative to consumer-facing generative AI is structural: vendors selling predictive maintenance systems sell to operations and engineering leaders, not marketing departments, so there's simply less incentive to publicise a deployment. There's also a competitive reason for manufacturers to stay quiet — a working predictive maintenance system is a genuine operational edge over rivals who haven't made the same investment, and there's little upside in advertising it.
The gap most worth watching now isn't between manufacturing and other sectors — it's between large manufacturers, who can absorb the upfront cost of sensors, data infrastructure and specialist staff, and smaller ones who often can't. Closing that gap without leaving smaller UK manufacturers behind is likely to be the more consequential story over the next few years than the headline adoption numbers themselves.