What should a window and door installer actually do about AI in 2026?

Why owners freeze on this

Because the market is loud and almost none of it is aimed at you.

You've had the LinkedIn posts, the cold emails, the agency that wanted £2k a month to "implement AI". Somewhere in there is something genuinely useful, but nobody selling it will tell you what it replaces, what it costs, or what happens when it goes wrong. So you do nothing, which is a rational response to bad information.

The second reason is scar tissue. Most installers have bought software before that nobody used. That experience teaches a lesson — systems don't stick here — which is the wrong lesson, but an understandable one.

What doing nothing costs

Not much this year. More every year after.

The window market isn't collapsing, but it is getting more competitive on service rather than price. FENSA's technical manager Sam Davies told Window News in April 2026 that the industry is seeing more cautious decision-making from homeowners, while noting glazing holds up better than other retrofit trades because demand comes from able-to-pay consumers rather than government schemes. Cautious buyers compare more, hesitate longer and are more sensitive to how a firm behaves.

There's a trust dimension too. A homeowner research report published July 2026, reported in Window News, found more than one in four homeowners hit a problem with a recent repair or improvement project; among those who lost money the average loss was around £750, and above £5,000 for roughly one in ten. Treat those as indicative, but the point stands: your prospect starts nervous. The firm that answers first, quotes fastest and communicates without being chased wins on the strength of that alone.

Your competitor down the road doesn't need to be better at fitting windows. He needs to answer the phone at half six.

Score yourself first

Five questions. Answer honestly, they take two minutes.

First, of the calls that came in last month, how many went unanswered? If you can't get that number from your phone provider in ten minutes, that's the answer. Second, what's the average gap between a survey and the priced quote landing, in days? Third, how many quotes sent in the last sixty days have had no follow-up contact at all? Fourth, can you say, right now, how many jobs are sold but not yet ordered, and what they're worth? Fifth, how much is outstanding on jobs where the fitters have already left site?

If four or five of those made you uncomfortable, you don't have an AI problem. You have four leaks and no gauge on any of them.

The order of operations

Fix them in this sequence, because each one makes the next one worth more.

First: answer the phone. Everything downstream depends on the enquiry existing. There is no point improving your quoting if a third of your enquiries never become conversations. This is where a voice agent earns its keep — not as a novelty, but as the thing that books a survey while you're on a scaffold. The honest version of what those can and can't do is in how many survey bookings a window firm loses to unanswered calls.

Second: get quotes out the same day. This is the highest-value single change available to most window firms, because homeowners gather two or three prices and the first one back, presented well, wins disproportionately. Details in why window and door quotes go cold.

Third: put the job somewhere everyone can see it. Once you're selling more, the operational mess costs more. One job record replacing the WhatsApp thread, the surveyor's phone and the whiteboard — covered in where a window installer should keep job information.

Fourth: collect the money automatically. More jobs fitted means more balances outstanding, and manual chasing collapses precisely when volume peaks. See how window installers stop losing final balances after the fit.

Doing these out of order is the common mistake. Automating invoice chasing while a third of enquiries go unanswered is polishing the back of the house while the front door is open.

How to tell useful from noise

Three questions to any vendor, including us.

What does this replace? If the answer is "nothing, it's additive", it's an extra thing for your team to maintain. Useful automation removes work.

Who runs it after it's built? A tool is a licence and a login. Most installers don't need another login — they need the thing configured, connected and operated. If the vendor hands you a platform and a training video, budget for the person who'll actually run it, because that person is currently you.

What happens when it gets something wrong? Any voice agent will occasionally mishandle a call. The question isn't whether, it's what the escalation looks like. If a vendor won't discuss failure modes, they haven't run one in production.

Where OrderWolf fits

We're the integrator and the operator, not a tool vendor. We rebuild the site to capture enquiries properly, run an AI receptionist so calls get answered around the clock, set up fast remote quoting, automate the follow-up, and put one CRM underneath the lot with live margin tracking. Then we run it. We don't sell tools — we run the whole stack.

With GEV, our Irish flagship client, the B2C close rate went from 45% to 70%, surveys booked rose 8%, and the team recovered roughly 2.5 admin hours per person per day, over a three-month deployment. Those are results from one installer, and outcomes depend on your starting point and your market.