The pattern in the companies that move fastest
The construction companies making progress with AI share three traits: leadership that understands how the work gets done, a single sharp problem they're trying to solve, and a refusal to buy systems they haven't seen working elsewhere.
They move fast because the people making the call understand the work AI is supposed to fix, not because they're tech-forward.
Why hands-on knowledge beats consultant slide decks
A consultant arriving at a £15m roofing business can describe the industry. They can talk in averages: "the typical roofer spends X hours on quoting." They can't tell you why a quote in your business takes longer than it should, because they haven't sat with your QS during the bid window.
A team close to the work knows two-thirds of the delay isn't producing the quote, it's chasing the structural engineer for one final number. AI can fix that, but only if someone diagnoses it specifically.
The advantage is specificity.
Three habits the best-run companies share
They write down what they want AI to do, in plain English, before any tool gets bought. One sentence per workflow. If they can't write it, they don't buy it.
They run small. The first AI system is one workflow, in one team, with one person responsible. A specific change to a specific thing.
They cancel quickly. If a tool isn't producing the result inside six weeks, they kill it. Carrying a non-working pilot costs more than admitting it didn't work.
What this means if you're a £10–50m company
Copying the pattern has nothing to do with company size or who owns the business. Spend a week alongside the workflow you're trying to change before you buy anything. Sit with the QS. Sit with the bid manager. Watch a job from enquiry to invoice.
The leadership teams behind high-performing AI deployments in construction have operational advantages, not technical ones. Any leader willing to do the work can have them.