The gap between ambition and reality is wide. Over 50% of manufacturers cite skills shortages as the main barrier to AI adoption, and only 2% say AI is widely embedded across their operations, according to Make UK's AI, Skills and the Future of UK Manufacturing report. So the technology is arriving, but the people who can direct it are the constraint.
That constraint is national, not local. During London Tech Week 2025, the UK government announced a partnership to upskill 7.5 million UK workers in essential AI skills by 2030, equivalent to around 20% of the UK workforce, techUK reported. When a fifth of the workforce needs new skills, hiring stops being a box-ticking exercise and becomes a judgement about where people add value that machines don't.
The entry routes are shifting too. Apprenticeships have risen from 3% of AI hires in 2020 to 19% in 2025, per the GOV.UK AI Labour Market Survey 2025. Plants are growing their own talent, not just buying it in, which tells you the skills they want are hard to find ready-made.
The direction of travel is clear. Six in ten (59%) manufacturers cited automation and half (50%) cited wider digitalisation as trends changing jobs and skills needs, and 74% expect demand for cognitive and meta-cognitive skills to increase by 2030, according to Make UK's 2030 Skills: Closing the Gap report. Cognitive and meta-cognitive skills are, in plain terms, thinking well and knowing how you think: judgement, problem-solving and learning on the job.
Those are exactly the things agents stop short of. An agent can schedule a line, flag a fault or draft a report, but it can't weigh a safety call against a deadline or read a situation no training data covered. That is the human layer plants are hiring for now.
Firms are also moving people rather than replacing them. 40% of manufacturers are looking to upskill their existing workforce, specifically for redeployment as a result of automation and digitalisation, and this figure increases for firms with more than 1,000 employees, the Make UK Executive Survey 2026 found. The pattern is consistent: automate the routine, redeploy the person to the judgement.
Start by writing the role around the decisions, not the tasks. If an agent will handle the routine, ask what calls a person still has to make when the data is messy, the clock is running and the standard answer doesn't fit. Those calls are your real job description.
Then test for them. Use work samples and scenario questions that put someone in a live trade-off, rather than screening only for tool familiarity, which dates fast. The skills that hold value are the cognitive and meta-cognitive ones Make UK flags as rising to 2030, so hire for how someone reasons, not just what they already know.
Finally, build a route in. With apprenticeships now a fifth of AI hires per the GOV.UK AI Labour Market Survey 2025, the plants hiring best are the ones growing judgement in-house as well as recruiting it. Keep the learning path open and your talent pipeline widens.
When you know the judgement you're hiring for, the task is finding the few people who have it. We search a database of 15 million candidates and rank a shortlist fast, so you spend your time assessing the decision-makers rather than sifting CVs. Our recruitment agent manages recruitment end to end for 8% on a successful hire, with no monthly fee and no upfront cost, through Reed.ai. That's end-to-end recruitment built around finding the right person. Tell us the judgement your plant needs, and we'll get started today.
