Why construction needs automation to tackle its labour shortage

Construction has a labour problem and the obvious solution to a labour problem is to hire more people. But with fewer people entering the construction industry than ever before, and a ticking timebomb of pending retirements, we can’t recruit our way out of this one. The question becomes how we build more with fewer hands – and that is what makes automation a necessity rather than a nice-to-have.

I came into construction from the outside, having trained as a criminal psychologist in a previous life, so I tend to read the numbers without the baggage of “this is how we’ve always done it.” On labour, the numbers are stark.

Around 35% of the UK construction workforce is over 50, and more than three quarters of a million people are expected to retire by 2036. CITB says we need roughly 240,000 extra workers over the next five years just to keep pace. Ireland is in the same place, with Property Industry Ireland putting the gap at up to 110,000. We’re trying to build more than we ever have, with a workforce that’s shrinking and getting older.

Recruitment, skills development and early career programmes remain essential. BAM UK & Ireland supports more than 300 apprentices on its early careers programme at any one time, but this is a demographic problem, and not just a recruitment blip. The people retiring this decade are in their 50s and 60s today, so the shortage is already baked in. Apprentices and graduates will be worth their weight in gold and we must continue to pour energy into them – but a trade or a degree takes years. Even a strong recruitment drive starting now won’t fix things this decade.

The construction industry already knows where this leads. In RICS’ most recent global productivity review, skilled worker availability came out as the single highest-impact factor on productivity, above technology and above sustainability. Skilled people are the bottleneck. You can buy a machine in a quarter. Developing skilled people takes years.

That is why we have to elevate what each person is able to deliver and technology is the only realistic lever. Not because technology replaces skilled people, but because it allows every skilled person to deliver more.

At BAM, any technology has to earn its place. On a remote Shetland site we used a laser-scanning robotic dog, controlled over 5G, to record more than 55,000 sq m of difficult terrain in weather you wouldn’t repeatedly send people into. That’s the model in miniature – give the hazardous, repetitive, hard-to-resource tasks to machines. Free skilled people for the work that needs judgement.

AI does the same thing to a different task: it helps skilled people do more. We’re currently experimenting with a tool that turns a routine photographic site walk into a digital comparison against the project plan. The AI tool maps the images to the plan, flags exactly where reality and programme have diverged, and lets a project manager step back in time to compare against earlier scans. If this can be proven and scaled, chasing subjective updates from across the site would be a thing of the past. It doesn’t replace the project manager. It hands them deep insight in a fraction of the time, so our experienced people spend time on decisions rather than data collection.

The hard part isn’t the technology. It’s the convergence between people who have been doing this job for 20 or 30 years and technology tools that are brand new. Those two things don’t naturally gel. Construction runs at high pressure with almost no slack to experiment and a project manager who has delivered tens of millions of pounds across multiple major schemes is understandably going to trust their gut and their experience over an unproven piece of tech. That instinct is a tough barrier to work around – and it’s an instinct that is often right.

Which is why the tech has to prove itself on their terms. The reason something like the progress comparison AI tool earns trust, is that it doesn’t ask anyone to take a leap of faith. It hands an experienced project team the objective evidence that backs or challenges what their gut already suspects, and it does it faster than they could gather it themselves. Adoption follows usefulness, not the other way round.

McKinsey estimates that global construction productivity grew about 0.4% a year between 2000 and 2022, against roughly 2% for the wider economy. We’ve stood still for two decades. Closing that gap means real investment in capability, as well as unglamorous spend on cleansing data, upskilling staff and closely managing the technology adoption. You don’t get the dividend without the groundwork and every pound of innovation in this sector is hard fought for.

This is not technology over people. The two run together: bring more people in and use technology to reduce how dependent each project is on human hours. Identify a real use case. Build it securely, prove the return, then scale. The aim isn’t simply to build faster; it’s to build safer, with greater consistency and better information.

The demographic numbers aren’t in dispute. The productivity numbers haven’t moved in 20 years. Put them together and the conclusion is hard to avoid. You can argue about the pace, but you can’t argue about the direction.