Measurement & Reporting · 9 min read · July 15, 2026

How many roofing enquiries come from AI recommendations? Numbers from the roofing trade

More and more homeowners ask ChatGPT, Gemini or Perplexity for a roofer before they even open Google. Yet almost no business knows how many renovation enquiries actually result from that. This guide shows you how to make the AI recommendation measurable in the roofing trade, which numbers are realistic and where the typical measurement mistakes lurk.

Why the AI recommendation suddenly matters in the roofing trade

Just three years ago, nearly every roof renovation followed the same path: the homeowner types "roofer near me" into Google, clicks one of the top results, gets two or three quotes. Today a new step slots in ahead of that. More and more people first ask ChatGPT or Perplexity things like "What does a new roof covering cost for a single-family home?" or "How do I recognise a reputable roofer?". The AI answers – and sometimes names specific businesses, criteria or regions.

For you as a roofer, that means: part of your future clientele makes their shortlist inside a chat window you never get to see. If the AI recommends competitors there, or simply doesn't know your business, you lose enquiries before they even arise. That is exactly why the question becomes so important: how many renovation enquiries already come through this AI recommendation – and how do you make that visible?

The honest truth up front: no one in the roofing trade can guarantee you an exact percentage today. But you can absolutely build a solid picture. It just takes the right measurement points instead of gut feeling.

What GEO concretely means for a roofing business

GEO stands for Generative Engine Optimization. It sounds clunky, but it means something down-to-earth: you make sure generative AI systems know your business, classify it correctly and recommend it at the right moment. It is the successor to classic SEO, only the stage is no longer the Google results list but the answer in the chat. For a roofer, good GEO means: when someone asks "Roof renovation on an old building – tiles or metal?", your name or your region appears in a sensible answer.

The difference from Google is decisive. On Google there are ten blue links and the user chooses for themselves. The AI often gives just one summarised answer with one to three named options. Whoever doesn't appear there practically doesn't exist for the person asking. Visibility thus shifts from a ranking to a yes-or-no. That makes every single mention far more valuable than a sixth place on Google.

For the roofing trade there is an added twist: your services are strongly local and strongly trust-based. A roof can't be ordered online. Here the AI recommends criteria and types of business rather than anonymous mass goods. That plays right into the hands of well-run specialist firms, provided they lay out their facts cleanly on the web.

The questions homeowners really ask the AI

Before you measure, you need to know what is actually being asked. From conversations with businesses and from typical search patterns, clear themes emerge. Cost questions dominate: "What does a roof renovation with insulation cost for 120 square metres?" or "Is above-rafter insulation worth it compared with between-rafter insulation?". People who ask questions like these are on the verge of a real investment – so exactly your target group.

Alongside them are trust and selection questions: "How do I recognise a reputable roofer?", "What questions should I ask before hiring one?", "How long does a complete re-roofing take?". And finally the technical decision questions: tiles versus concrete roof tiles, adding photovoltaics to the roof at the same time, repair or complete re-covering after storm damage.

The point is: for almost all of these questions the AI can bring a regional specialist firm into play – or name criteria that fit you exactly. If you know which questions your customers ask, you can test these questions yourself and see whether and how your business appears in the answers.

How to measure the AI recommendation – four practical routes

The first route is asking your customers directly. Add a simple line to your enquiry form and your phone script: "How did you hear about us?" with the extra answer option "via ChatGPT, Gemini or another AI". It sounds trivial, but it is the most honest source. Many businesses discover this way, for the first time ever, that AI recommendations are already bringing them enquiries.

The second route is your own test. Ask the AI systems the typical questions of your region yourself, for example "Good roofer for roof renovation in the Rosenheim area". Note whether your business appears, which competitors get named and which facts the AI states about you. Repeat this monthly and you'll see development instead of a snapshot.

The third route is your web server. Access statistics now show AI crawlers and referral sources such as chatgpt.com or perplexity.ai. A glance at the logs or your analytics tool tells you whether visitors from AI answers are clicking through to your site at all. The fourth route, finally, is a specialised monitoring tool that automatically checks how often and in what context your business gets named in AI answers.

Which numbers are realistic in the roofing trade right now

Now for the honest framing, because exaggeration doesn't help you. At most regional roofing businesses, the share of enquiries that come directly and consciously via an AI recommendation is still in the single-digit percentages today – often somewhere between two and eight percent of first contacts. That sounds like little. But it is a figure that stood at practically zero two years ago.

More important than the current absolute value is the direction of the curve. Use of generative AI is growing fast, and especially for larger, expensive decisions like a roof renovation, people research thoroughly beforehand. On top of that comes a dark figure: many customers say on the phone "I found you on the internet", even though the AI was the actual trigger. So the real influence is greater than what gets directly measured.

Run the numbers for yourself once. If you get 300 enquiries a year and only five percent are AI-driven, that's 15 enquiries. With an average renovation in the five-figure range and a solid closing rate, that is an order volume where paying attention has long been worthwhile.

The typical measurement mistakes – and how to avoid them

The most common mistake is all-or-nothing attribution. A customer names "Google" as the source on the phone, so you book them as a Google enquiry. Yet they may have asked ChatGPT beforehand, remembered your name and then googled you deliberately. The AI was the first impulse, but it doesn't appear in your statistics at all. That is why the open question "What first brought you to us?" is more valuable than a fixed selection list.

The second mistake is the one-off test. You ask ChatGPT once, your business doesn't come up, and you conclude that AI brings nothing. But AI answers fluctuate depending on wording, location context and model version. A single query is statistically worthless. Only repeated tests over weeks, with varying phrasings, produce a reliable picture.

The third mistake is looking only at clicks. Unlike on Google, with AI often no one clicks a link – the answer in the chat is enough. The customer remembers your name and calls directly. Anyone who measures AI success only by website clicks massively underestimates it. So measure the mention itself, not just the traffic.

What you can do so the AI recommends you more often

Generative AI draws its answers from what is said about you on the web. The first lever is therefore clean, consistent facts: the same business name, the same address, the same phone number and the same range of services on your website, in trade directories, in your Google Business Profile and on review portals. Contradictions unsettle the AI, and uncertain systems prefer to recommend the competitor with a clear data picture.

The second lever is real content that answers exactly the questions your customers have. A guide on your site about "Between-rafter or above-rafter insulation – what suits my roof?" or "Storm damage to the roof: the first steps" gives the AI usable material. Write the way your customers ask, with clear answers, concrete numbers and regional reference. Passages exactly like these are what the AI systems like to quote.

The third lever is reviews and mentions by third parties. When customers recommend you in forums, on review platforms or on local portals, that anchors your name in the AI's training and search material. Actively ask satisfied customers for an honest review. That helps not only on Google but increasingly with the AI recommendation too.

A realistic roadmap for the next three months

Start small and concrete. Month one: add the origin question including the AI option to your enquiry form and your phone script, and put together a simple list of ten typical customer questions from your region. Test these questions once with ChatGPT and Perplexity and record whether and how you appear. That is your baseline measurement, your starting point.

Month two: tidy up your facts. Check name, address and services across all platforms and eliminate contradictions. Write two guide articles on the most frequent customer questions, for example the cost and process of a roof renovation. In parallel, ask five satisfied customers for a review. This work pays into both Google and the AI recommendation at the same time.

Month three: repeat your tests and compare with the baseline. Are you appearing more often now? Do customers name AI as a source more frequently on the phone? This is how gut feeling turns into a real, recurring measurement. And you make decisions about your visibility where your customers will look first in future – not only once the enquiries have already landed with the competitor.

Common questions

How do I tell whether a renovation enquiry came via an AI recommendation?

Most reliably through the direct question at first contact: "What first brought you to us?" with the AI option. Additionally, your web server logs show referral sources such as chatgpt.com or perplexity.ai. But reckon with a dark figure, because many customers still name "Google" or "the internet" as the source after the AI impulse.

Is GEO even worth it yet for a small regional roofing business?

Yes, precisely because the competition here is still thin. Most roofing businesses aren't even bothering to lay out their facts cleanly. The effort is manageable: consistent business data, a few real guide contents answering customer questions, and active reviews. This pays into Google and the AI recommendation at the same time, so you lose nothing.

Why does ChatGPT name my business today but not tomorrow?

AI answers fluctuate depending on wording, location context and model version. A single query is therefore never meaningful. Test the same questions repeatedly over several weeks and with slightly different phrasings. Only this pattern shows whether your business is mentioned stably or just turned up by chance. Consistent facts on the web raise the hit rate considerably.

Read on

Content & Answer Pages