TL;DR: Contractors have been getting less worried about AI taking construction jobs every year for three years, at exactly the point the technology got good. Meanwhile the share of firms struggling to hire AI specialists jumped from 30% to 47% in a single year. The displacement is not arriving. A different shortage is.
The people who would lose their jobs to AI have looked at it and concluded they won't.
That is not optimism from a technology vendor. It is what contractors told the Associated General Contractors of America and NCCER in their 2025 Workforce Survey, which drew responses from roughly 1,400 firms. Asked how robotics and artificial intelligence would affect construction jobs over the next five years, 12% expected AI to eliminate jobs.
The same question produced 14% in 2024 and 17% in 2023.
Fear of displacement fell for three consecutive years across the exact period when the underlying technology improved most. Either the industry is asleep, or the people doing the work understand something that the coverage does not.
What do contractors expect instead?
Automation of the parts of the job nobody wanted.
In the same survey, 45% expected robotics and AI to affect construction jobs positively by automating manual, error-prone tasks, and 44% expected improved job quality with safer and more productive workers. Roughly a third expected no significant effect at all.
Note what is missing from those options: mass substitution. The modal contractor expectation is that machines take the repetitive and dangerous work and the crew stays.
Dodge Construction Network's survey of 235 contractors, run with CMiC in September and October 2025, points the same way. 85% expected AI to reduce time spent on repetitive tasks. Not headcount. Time on repetition.
Doesn't the robotics evidence say otherwise?
This is the strongest argument against my position, so let me make it properly rather than waving at it.
Construction robotics is genuinely working now, in specific scopes, with real numbers. The clearest case is rebar tying. Advanced Construction Robotics states its TyBOT achieves an active tie rate of over 1,200 ties per hour and more than 9,600 ties in a standard eight-hour shift, in all weather, day or night. Its published project results include a 30% schedule saving on the I-30 Crossing in Arkansas, and on the Koppel Bridge in Pennsylvania a 50% productivity increase with customer crew manhours down 34%. On the Port St. Lucie Bridge, productivity up 59% and crew manhours down 37%.
Those are vendor figures and should be read as such. They are also specific, project-named and checkable, which is more than most automation claims offer.
So a machine really does displace roughly a third of the crew hours on that scope. Anyone arguing AI changes nothing in construction is not reading.
Why doesn't that generalise?
Because of what the same specification says about scope.
TyBOT self-navigates and ties bulk horizontal rebar intersections, up to #8 by #9 bar, across expandable widths from 10 to 117 feet.
Bulk. Horizontal. Intersections. That is a flat mat on a bridge deck or an industrial slab, which is one of the most repetitive, most geometrically regular tasks on any site. It is precisely the task you would automate first, and after years of development it is where the technology has landed.
It does not do congested vertical steel. It does not work around embeds and penetrations in a tight footprint. It does not tie the awkward third of the job that takes an ironworker's judgment.
And the case studies report crew manhours falling by around a third, not to zero. The crew is still there, doing the part the machine cannot. That is the shape of the whole category: human-robot teaming on narrow scopes, not substitution.
So where is the actual labour crisis?
In the people who would deploy any of this.
The AGC survey found that among salaried positions, difficulty filling AI personnel and specialist roles was reported by 47% of respondents, up from 30% in 2024. That was the largest single-year jump in the survey.
Set that beside the craft picture from the same source. 88% of firms employing craft workers had openings. 57% said available candidates were not qualified. Worker shortages remained the most cited cause of project delays at 45% of respondents, though that has fallen every year from 66% in 2022.
Put those together and the industry's position is uncomfortable in a way the displacement narrative completely misses. Construction is short of electricians and ironworkers, and it is now also short of the people who would implement the technology sold as the answer to being short of electricians and ironworkers.
Any vendor pitching AI as a workforce fix is proposing a solution that depends on a scarcer hire than the one you already cannot make.
What does this mean for the person holding the tools?
Your job is not going away. The paperwork around it might, and that is the good news you should actually want.
The tasks realistically exposed over the next five years are the ones already half-automated in every other industry: retyping numbers off documents, chasing signatures, assembling the same report weekly, finding a photo from four months ago. None of that is why anyone became a tradesperson.
The tasks that stay are the ones requiring judgment on an uneven site with incomplete information, which is most of the job and the part that resists specification.
The genuine risk is not replacement. It is that the accumulated judgment walking out the door with retirements does not get transferred, and AI is not close to substituting for it. The 57% figure on unqualified candidates is a training problem, and no model fixes a training problem.
The prediction, stated so it can be wrong
I think the AGC's job-elimination number keeps falling or stays flat through 2028, and that the AI specialist hiring difficulty figure rises above 50% before it peaks.
Concretely: if the share of contractors expecting AI to eliminate construction jobs exceeds 17% in any AGC workforce survey through 2028, returning to its 2023 level, I was wrong about the direction. And if AI specialist hiring difficulty falls back below 35%, then either the tooling got simple enough not to need specialists, or contractors gave up on it. Both would mean I read this wrong.
What I would not bet against is the robotics case in narrow scopes. Rebar tying worked. Layout worked. Something else regular and repetitive will work next. That is displacement of tasks, on schedules measured in years, in the specific corners where geometry cooperates.
The thing that will not happen on any timeline worth planning around is a jobsite that builds without a crew.
Frequently Asked Questions
Should a young person still enter the trades?
On this evidence, yes, and the numbers are unusually clear about it. 88% of firms report craft openings and 57% say the available candidates are not qualified, which describes a market where a credentialed, competent tradesperson has substantial bargaining power. The occupations most exposed to automation are administrative, not manual-with-judgment.
What AI skills should I be hiring for?
Not model building. The scarce capability is someone who understands your operation well enough to specify what a tool should do, and is technical enough to hold a vendor to it. That is closer to a good project manager who is comfortable with software than to a data scientist, and it is a role you can often develop internally faster than you can recruit it.
Does AI help me hire better workers?
Marginally and mostly in screening logistics rather than judgment. It can sort applications and schedule faster, which matters when your problem is response time. It cannot assess whether someone will show up, which the AGC survey identifies as a leading candidate problem at 48% of firms reporting new hires failing to appear or quickly quitting.
Is any trade genuinely at risk?
The most exposed work is repetitive, geometrically regular, and performed in controlled conditions, which describes prefabrication and shop work more than field installation. If your work happens in a factory rather than on an uneven site, the automation timeline is shorter, and that has been true since long before AI.
Why do vendors push the displacement story if contractors do not believe it?
Because urgency sells and fear is the fastest route to urgency. It is worth noticing that the displacement narrative is most enthusiastically promoted by people selling the thing that supposedly does the displacing.
If someone is selling you AI on the basis that it solves your labour shortage, ask which scope, and ask how many people the crew still needs. The honest answers are narrower and more useful than the pitch.
See ClickWerxs AI services or get in touch. For the adoption picture behind this argument, what AI actually does for a construction business covers the survey data in full, and how to evaluate an AI vendor covers the questions to ask.
Sources
- Associated General Contractors of America and NCCER, 2025 Workforce Survey Analysis — approximately 1,400 firms; 12% expect robotics and AI to eliminate construction jobs, against 14% in 2024 and 17% in 2023; 45% expect positive impact through automating manual, error-prone tasks; 44% expect improved job quality and safer, more productive workers; about one third expect no significant effect; difficulty filling AI personnel and specialist positions reported by 47%, up from 30% in 2024; 88% of firms employing craft workers report openings; 57% report available candidates not qualified; 48% report new hires failing to appear or quickly quitting; 45% report project delays from worker shortages, against 54% in 2024, 61% in 2023 and 66% in 2022. agc.org
- Advanced Construction Robotics, TyBOT product specification — manufacturer states an active tie rate of 1,200+ ties per hour and 9,600+ ties per standard eight-hour shift; ties bulk horizontal rebar intersections up to #8x#9 bar at expandable widths of 10 to 117 feet; minimum 25% schedule savings claimed, with published project results including 30% schedule savings at the I-30 Crossing (Arkansas), 50% productivity increase and 34% reduction in customer crew manhours at Koppel Bridge (Pennsylvania), and 59% productivity increase with 37% reduction in crew manhours at Port St. Lucie Bridge (Florida). Vendor-stated figures. constructionrobots.com
- Dodge Construction Network with CMiC — 235 general and trade contractors surveyed September and October 2025; 85% expect AI to reduce time spent on repetitive tasks. Reported via constructiondive.com
This post reflects operator opinion and is not legal, financial, or professional advice. Third-party survey figures are attributed with sample sizes and dates; robotics performance figures are manufacturer claims and are identified as such. ClickWerxs sells AI implementation services and earns revenue from those engagements. Predictions in this post are the author's and are stated in falsifiable terms deliberately; they are not forecasts you should plan a business on.
Kaleb Dickhaut — Founder, ClickWerxs. Kaleb built ClickWerxs from the ground up, from payment processing ISO to the Command Center platform to the AI SEO methodology the blog runs on. He has onboarded hundreds of small businesses onto payment and CRM systems. linkedin.com/in/kaleb-dickhaut