Industrial employers sitting on unfilled technical roles and a generation locked out of work are two halves of one fixable equation.
The latest figures from the Office of National Statistics reveal that around 13% of UK 16-24 years olds are not in employment, education or training. The government’s own interim review has warned this could mark the makings of a “lost generation.”
At the same time, close to a fifth of the engineering and manufacturing workforce is expected to retire by the end of the year, taking with it an estimated half a million skilled engineers, technicians and operators.
Latest Videos FromTechRadar
Chief Product & Customer Officer at IFS.
One side of the ledger is a generation that wants to work and cannot get in. The other is an industrial base that needs people and is about to lose its most experienced. Treated separately, both look intractable. Treated together, they start to look like a single problem with a single answer. The solution: training AI agents to become autonomous and fill this gap.
A knowledge problem, not a hiring problem
The first mistake is to see the industrial shortage as a hiring problem. Recruiters and trade bodies have been clear that external hiring alone cannot close the gap: the issue is not just empty seats but the decades of judgment walking out the door with every retirement. When a maintenance lead or a supplier audit manager leaves, the loss is rarely the task itself. It is the accumulated reasoning behind it, and the why that never made it into a manual.
That is precisely why hiring and knowledge transfer demand completely different solutions. A vacancy can be filled in months. Thirty years of tacit knowledge cannot be replicated by recruitment, and the so-called “hollowed-out middle” – the shortage of workers with five to fifteen years’ experience – means there is often no intermediate layer left to pass that knowledge down. The conveyor belt that once carried expertise from veteran to apprentice has lost its middle section.
Where AI changes the arithmetic
This is where AI changes the arithmetic, and not in the way the headlines assume. The breakthrough is not a chatbot that answers questions. It is the ability to capture the relationship between an organization’s structured records, the purchase orders, the part numbers, and the unstructured reasoning that surrounds them: the emails, the annotations, the informal decisions made in the margins.
Map how the two influence each other, embed that context into the daily workflow, and decades of institutional memory become an operational asset rather than a folder no one can find.
Done well, this compresses the experience curve. Instead of acquiring judgment slowly, over years, a newer engineer inherits a working version of it from their first day on the job. The relevant history, the prior decisions, and the reasoning surfaced from the very beginning. The point is not that experience stops mattering. It is that the ramp becomes far shorter.
The window to exchange knowledge is narrow, and it is closing now. Context can only be captured while the people who hold it are still in the building. Once an expert retires, the structured data remains, but the reasoning that made it usable leaves with them. Every quarter of delay is a quarter of irreplaceable judgment lost. The retirement wave is not a distant problem to be scheduled around; it is already underway.
Where AI is used the wrong way
There is an honest objection to confront. AI is widely blamed for hollowing out entry-level work, and the data gives the charge weight: the same interim review found that entry-level roles have become less plentiful and more demanding, recruitment more remote, automated and less human, and apprenticeship starts for young people are down by more than 40%.
But AI is rarely the real culprit; the mindset behind it is. Too many organizations are still buying the technology the way they bought software a decade ago – as a line item that removes people rather than a capability that makes them more valuable. Pointed at entry-level work, AI will cut it and call the result “efficiency”.
Pointed at the opposite goal, making inexperienced people productive faster, the same technology widens the door rather than narrows it. The real question is not whether AI has to close off the old way in, but whether companies choose to use it to open a new one.
A different way in
This only works if the argument is honest about who these young people are. The cohort behind the NEET numbers is not, for the most part, trained engineers waiting for a vacancy. Six in ten have never held a job at all; many are contending with poor health, disrupted schooling or years of disengagement. Employers are willing to take them on yet hesitate because an unproven candidate feels like a risk a smaller firm cannot afford.
AI lowers the cost of that bet. When a system surfaces the right context and lets a newcomer direct complex processes in plain language, the absence of a track record stops being disqualifier. For a 23-year-old walking onto a factory floor for the first time, that is the difference between being screened out and being let in.
That is the opportunity hiding inside the two statistics. The retirement cliff and the NEET crisis are usually filed under different departments: workforce planning on one side, social policy on the other. They belong on the same page. Capture the expertise that is about to leave, put it within reach of the people who are desperate to start, and the industrial sector does not merely plug a gap. It builds the on-ramp for a generation that has been told, too often, that there was no way in.
We’ve featured the best AI tool.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit
