Trunk Tools CEO Dr. Sarah Buchner says the era of AI early adopters in construction ended in 2023. Piloting defined 2025. In 2026, billion-dollar contractors are signing full enterprise rollouts, and the firms still “waiting to see” are now late adopters carrying real competitive risk. Even hourly-billing consultancies are adopting tools that cut the hours they charge for, reasoning it’s better to drive the race than get run over by it.
A construction executive recently told Sarah Buchner that his company didn’t want to be an early adopter of AI. Her response, shared on the latest episode of Bricks, Bucks & Bytes, was blunt: “What are you talking about? Early adopters was in 2023. There are no more early adopters.”
Buchner has a useful vantage point for this claim. Her company, Trunk Tools, closed a $40 million Series B led by Insight Partners in July 2025, taking total funding to $70 million, and she spends her weeks negotiating with the C-suites of some of the largest general contractors in the US. What she described on the show is a distinct shift in how construction buys AI, and it happened fast.
If you run a contracting business, the useful question is no longer whether AI works. It’s where your firm sits on a buying curve that has already moved three times.
The three phases of buying AI in construction
From scary pilots to enterprise negotiations in three years
Buchner sketched the timeline from her seat as a vendor. Two to three years ago, even a pilot felt risky to most contractors, and only genuine early adopters jumped in. Then came 2025, which she summed up as “pilot, pilot, pilot.” Nearly everyone was testing something, but very few had the nerve to scale. The exceptions were firms like Gilbane and Suffolk, which ENR notes were among the earliest to put natural-language AI on jobsites, while the average GC stayed in test mode.
This year is different. “We are having I don’t know how many full enterprise rollouts, full enterprise negotiations,” Buchner said. “We’re talking large, large agreements, because the C-suites of these companies understood that the change is happening.”
Her framing of the choice facing leadership teams was stark: hide and risk the whole future of your company, or jump on it. The shift from fear to commitment, in her telling, happened over roughly the last three months, and it’s most visible at billion-dollar-plus general contractors. We’ve tracked the same pattern from the buyer side in our piece on why customer adoption pace, not model capability, is the real brake on ConTech AI.
Waiting for the “best” tool is the wrong strategy
Train your people now, swap the software later
The counterintuitive part of Buchner’s argument is that picking the winning vendor matters less than most executives think. Agentic interfaces are simple by design, and most of the useful agents run in the background. Her advice to customers: commit to a platform now, train everyone, and if a better provider exists in two years, “you can switch within a day because it’s all the same interface.”
What can’t be swapped in a day is a workforce that knows how to think and work with AI. That’s the gap she sees opening in the market right now, between firms training their people today and firms waiting to see who wins the tech race. The waiting firms lose twice: they miss the productivity gains, and their people start from zero when they finally move.
This is probably the largest economic change that we will see in our lifetime, at least so far. And it means a mindset shift. Mindset shift in our industry is way harder than swapping a tech tool in and out.Dr. Sarah Buchner, CEO, Trunk Tools
Even the hourly billers are on board
When consultancies cut their own billable hours, the shift is real
The most surprising signal Buchner shared involves services firms: forensic engineering practices, owner’s representatives, businesses that bill by the hour. Trunk Tools is now selling them software that reduces the hours they charge clients for. On paper, that looks like commercial self-harm.
When she asked one private-equity-backed customer why they’d do it, the answer stuck with her: “It’s going to be a race to the bottom, so we might as well be the ones that are driving the race.” Her read is that they’re right. Clients won’t keep paying hourly rates for white-collar work an agent can handle most of, so the smart firms are rebuilding their pricing and value propositions before the market forces them to.
Get the AI adoption signals before your competitors do
One sharp briefing each week on the deals, tools, and shifts reshaping construction.
Join 3000+ ReadersWhat AI efficiency actually means in a labor-starved industry
Revenue per head, not headcount reduction
Buchner pushed back on the idea that AI adoption in construction leads to layoffs. The industry has a labor shortage and a retirement wave, so the goal is different: increase revenue and EBITDA per employee. “Nobody needs to lay anybody off in construction,” she said. A GC that wants to ride the data center boom can’t add 50% more headcount, because those people don’t exist. The only route to that growth is making each existing employee more productive.
She does expect white-collar workloads to change substantially over the next two to three years, the way they already have in software engineering and law. Jobs won’t disappear so much as stop looking like they do today. For an industry where 78% of trade contractors already use AI in some form, that transition is less hypothetical than the headlines suggest.
Here’s the buying curve as Buchner described it, and what each phase signals about a firm’s position now:
| Phase | Who moved | Typical commitment | What it means in 2026 |
|---|---|---|---|
| 2023: Early adopters | A handful of genuinely experimental GCs | Small pilots, single projects | These firms now have 3 years of AI-trained staff |
| 2024-2025: Pilot era | Almost everyone, led by firms like Gilbane and Suffolk at scale | Pilots everywhere, scaling rare | Piloting alone no longer differentiates |
| 2026: Enterprise commitment | Billion-dollar contractors, plus hourly-billing services firms | Full enterprise rollouts and multi-year agreements | The current competitive baseline |
| Still waiting | Firms that “don’t want to be early adopters” | None | Now mid-to-late adopters, with a growing skills gap |
Dr. Sarah Buchner is the founder and CEO of Trunk Tools, a New York-based construction AI company. She started her career on construction sites before moving into technology, and her company raised a $40 million Series B led by Insight Partners in July 2025, bringing total funding to $70 million. Her view of the market comes from active enterprise negotiations with large US general contractors. (Source)
She means the early-adopter window closed around 2023. Firms adopting AI now aren’t taking a pioneering risk; they’re catching up to a market where pilots are routine and enterprise rollouts are underway. A company that starts today is a mid-to-late adopter by her definition, whatever it prefers to call itself. (Source)
Buchner named Gilbane and Suffolk as the firms that moved beyond pilots while the average GC was still testing. ENR’s reporting on Trunk Tools confirms both contractors were among the earliest adopters of its natural-language AI on jobsites, starting back in 2023. (Source)
Because the alternative is worse. Once agents can handle most of a white-collar workflow, clients will stop paying full hourly rates for it regardless of what any individual firm does. The services firms Buchner described, many private-equity backed, would rather lead that repricing and design new business models on their own terms than defend a dying one. (Source)
Not in Buchner’s view. Construction faces a labor shortage and a retirement wave, so the economic goal is higher revenue per employee, not fewer employees. She does expect the shape of white-collar construction roles to change considerably over the next two to three years, similar to what has already happened in software and law. (Source)
Because the durable asset is your trained workforce, not the software license. Buchner argues that agentic interfaces are simple enough that switching vendors later takes a day, while building a team that works fluently with AI takes years. Firms that delay both the tool and the training fall behind on the part that’s hard to recover. (Source)
Related Articles
“You Can Only Go As Fast As Your Customers”: The Quiet Brake on ConTech AI
78% of Trade Contractors Use AI: Adoption Is Way Ahead of What You Think
