AI voice for glaziers: quotes, bookings and board-ups

A glazing firm lives and dies by the phone. A smashed shopfront at 2am, a cracked patio door before a family holiday, an insurance-referred replacement that has to be surveyed this week: almost every job starts as an inbound call, and a large share of them arrive when the glaziers are up a ladder, mid board-up, or asleep. The UK doors and windows market is worth around £7.4 billion in 2025 and is forecast to reach £8.6 billion by 2030, a compound annual growth rate of roughly 2.7%. The demand is there. The bottleneck is answering it.

The wider trades picture makes the point sharper. A 2025 study of 142 UK small and medium businesses found that 47% of first-time calls went unanswered, and callers rarely wait: most ring the next glazier on the list rather than leave a voicemail. For a firm whose work sits inside a repair, maintenance and improvement sector worth more than £41 billion a year, every unanswered emergency board-up is a lost job and, often, a lost insurance relationship behind it.

This is not an argument for a bigger call centre. It is an argument for a voice line that never sleeps, triages correctly, and hands clean job data to the diary. That is a narrow, winnable use of AI, and it is exactly where a glazing business should start.

This guide is shipped by the team behind Dilr Voice, enterprise voice AI built for regulated and field-service deployments. Or see DATS, our five-stage AI consulting system.

What does an AI voice agent actually do for a glazing company?

An AI voice agent answers a glazing firm's inbound calls, works out whether the caller has an emergency board-up, an insurance-referred replacement, a survey request or a routine quote, captures the job address and the glass details, and books the right slot in the diary. Dilr Voice handles this on the first ring, day or night, and escalates anything genuinely urgent or advisory to a human. It is a booking and triage line, not a salesperson.

The reason this matters more for glaziers than for most trades is the mix of urgency. A boiler that stops working is uncomfortable; a broken ground-floor window is a security and safety problem that a customer wants solved tonight. The AI voice line's first job is to separate the "make it safe now" calls from the "book me a quote" calls, because those two go to completely different parts of the operation. Our boiler and heating-engineer booking guide covers the heating-trade version of the same problem, but the glazing urgency model is its own thing.

Adoption pressure is real, and the value gap is the point. McKinsey's State of AI, published in November 2025, found that around 88% of organisations now use AI in at least one function, yet only about 6% are mature enough to see material profit impact. Most firms are using AI without capturing value from it. For a glazing business, the phone line is the one place where a narrow, well-scoped agent turns adoption into booked revenue, which is the whole thesis behind our AI operating model consulting.

How does AI voice triage an emergency board-up from a routine quote?

The triage runs on the first two questions: is this an emergency, and is it insured. An emergency board-up (a smashed shopfront, a broken ground-floor window, storm damage) needs a make-safe visit against a response-time target, so the agent captures the address, the risk, and access details, then books the on-call slot. A routine quote (a misted double-glazed unit, a planned window replacement) goes to the survey diary. Dilr Voice keeps those two lanes separate from the opening exchange.

Getting the classification right at the start is what protects the response-time promise. If an emergency board-up is mislabelled as a quote, it sits in a survey queue while the customer stands next to a broken window; if a routine quote is treated as an emergency, an on-call glazier is pulled off higher-value work. This is the same first-move discipline we apply in our AI operating model consulting: decide the branch before you gather the detail. It is also why a generic reception script fails glaziers, and why the facilities-management triage pattern needs adapting rather than copying.

Can AI voice handle insurance-referred glazing work safely?

Yes, within a clear boundary. A large share of replacement glazing is insurance-referred, so the agent's job is to capture the policy number, the claim reference and the loss details, then route the caller into the insurer's or the firm's approved-supplier workflow. What it must not do is advise on cover, liability or settlement. Insurance distribution and claims handling are regulated activities under the Financial Services and Markets Act 2000, so Dilr Voice is scoped to intake and routing only.

That boundary is not a limitation, it is the design. The Financial Conduct Authority's perimeter guidance in PERG 5 treats acting as an intermediary in insurance, and handling claims, as activities that need authorisation or an appointed-representative arrangement. A glazing firm's voice line stays firmly on the operational side: it confirms the job, the reference and the access window, and it books the survey. Any question about whether the policy pays goes to the insurer or the broker, never to the agent. We scope this boundary explicitly in every deployment, which is part of what an AI execution office is for.

How does an AI voice line capture the right glass specification?

It captures glass detail by asking the questions a surveyor would ask, in the order a glazier needs them. Frame type (uPVC, aluminium or timber), single or double glazed, approximate pane size, whether the unit is misted or cracked, and where it sits in the property all shape the quote and the survey slot. Dilr Voice records these against the job and flags anything pointing to safety glazing, so the surveyor arrives ready rather than blind.

Context matters here: the vast majority of UK homes are already double glazed, so most glazing calls concern an existing opening with a known frame, uPVC, aluminium or timber, which is exactly the kind of structured, repeatable intake a voice agent does well. The agent does not measure, and it does not commit to a firm price. It gathers the spec cleanly and books the survey, which is where the accurate number gets set. This is the same intake discipline behind our physiotherapy clinic booking work, applied to glass.

The technical plumbing behind this is unglamorous but decisive. The voice layer connects to telephony through a provider such as Twilio, writes the captured job into the firm's diary or CRM, and sends an SMS confirmation with the reference. If you want to see the agent side of that stack, Dilr Voice is the product we build it on.

Where do FENSA, CERTASS and safety glazing rules fit in?

They set the compliance frame that the intake has to respect. FENSA and CERTASS are government-authorised competent person schemes: a registered installer can self-certify that replacement windows and external doors in existing dwellings meet the Building Regulations and register the work with the local authority, rather than the homeowner applying to Building Control. That covers replacement, in England and Wales. It does not cover an emergency board-up, and the voice agent should never imply otherwise.

Safety glazing is the part a glazing firm cannot get wrong, and it is the vocabulary the agent should recognise. Under Requirement K4 of Schedule 1 to the Building Regulations 2010, glazing that people are "likely to come into contact" with "while moving in or about the building" must, if broken on impact, "break in a way which is unlikely to cause injury", or "resist impact without breaking", or "be shielded or protected from impact". Approved Document K treats glazing within 800mm of floor level, and glazing in doors and side panels up to 1500mm, as critical locations that need safety glass to Class 3 of BS EN 12600 or Class C of BS 6206.

The practical point for the phone line: when a caller describes a broken pane in a door, a low-level window or a glazed side panel, that is a critical-location signal, and the agent should capture it so the surveyor quotes the correct safety glass. The Glass and Glazing Federation, whose members represent more than 60% of the industry's turnover, publishes the sector standards behind this. The agent is not making the compliance decision. It is making sure the human never surveys blind. Where captured call data touches personal information, the same discipline we set out in our privacy-notice and transparency guide applies.

How much does missed-call handling actually cost a glazing firm?

More than most owners think, because the loss compounds. A missed emergency call is not one lost job; it is the board-up, the follow-on replacement, and the insurer relationship that would have sent the next referrals. With nearly half of first-time calls to UK SMEs going unanswered and callers moving straight to a competitor, a single-van glazier can lose tens of thousands in pipeline a year to voicemail. An always-on voice line closes that leak without adding headcount.

The value framing beats a feature list every time. The question is not "does it have voicemail transcription", it is "what is a captured emergency board-up worth, times the ones you currently miss". That is the calculation our diagnostic-first approach exists to run before any deployment. The same near-universal-adoption, rare-value-capture gap that McKinsey measured across the economy shows up in miniature on a glazing firm's phone: the technology is available to everyone, but the booked-job outcome only accrues to the firm that scopes it to the real bottleneck.

The same operators who scope the call flow also run it: our AI execution office takes a fixed-fee mandate to deploy and govern the agent after go-live, so a glazing firm knows the payback before it signs.

What is the best AI voice setup for a UK glazing business in 2026?

The best setup is the one scoped to your booking bottleneck, not the one with the most features. For a glazing firm, that means an agent that triages emergency board-ups against a response-time target, captures insurance references without straying into advice, records glass spec and critical-location signals for the surveyor, and writes clean jobs into your existing diary. Dilr Voice is built for exactly this field-service pattern, but honesty demands naming where others fit.

Developer-first platforms such as Vapi, Retell AI and Synthflow are strong if you have an in-house engineer who wants to assemble and maintain the flow yourself, and PolyAI suits very high call volumes at large contact centres. If you are a glazing firm that wants the booking outcome rather than a build project, a managed deployment wins, because the value is in the triage logic and the diary integration, not the raw voice. There is a real scenario where a competitor is the right call: a single-location shopfitter with a developer on staff and simple, non-urgent enquiries may not need a managed service at all. Most glaziers, juggling emergencies and insurance work across a van fleet, do. We set out the trade-offs plainly in our approach and across the industries section of this blog.

Does AI voice work out of hours for emergency glazing?

Yes, and out of hours is where it earns its keep. Emergency glazing calls cluster in the evenings, at weekends and during storms, precisely when a small firm cannot staff a phone. An AI voice agent answers every one, captures the make-safe details, and either books the on-call glazier or escalates a genuine risk. The membership and renewals booking pattern shows the same always-on principle in a different sector.

Can it book against an engineer's route and diary?

It can, and that is the difference between a message-taker and a booking line. The agent checks availability in the firm's diary, respects the on-call rota, and books the slot with the address and job type attached, so the glazier's route is set before the call ends. Integrations to scheduling and CRM tools, often via Dilr Voice, are what turn a captured call into a routed job rather than a note someone has to re-key.

Is caller data handled in line with UK GDPR?

It should be, by design. A glazing call captures a name, an address and sometimes an insurance reference, all personal data, so lawful basis, a clear privacy notice and sensible retention are part of the build, not an afterthought. Where a caller needs to be identified before details are changed, the controls in our caller identity verification guide apply. The Information Commissioner's Office expects data minimisation and transparency, and a well-built agent captures only what the job needs.

Want to see this in production? Try Dilr Voice live, book an AI placement diagnostic, see our DATS methodology, or read about our approach to placing AI inside field-service operations.

Written by the Dilr.ai engineering team, practitioners who ship enterprise AI in production. We also help firms scope where AI belongs through our DATS methodology and can tell you who we are first. Follow us on LinkedIn for shipping notes, or subscribe via the RSS feed.