gptagency.io

Local & Industries · 9 min read · July 15, 2026

From Google profile to expert text: which sources the AI reads about your roofing business

When a homeowner asks the AI "Which roofer near me can renovate a thatched roof?", the model doesn't pull a secret database. It reads your Google profile, business directories, reviews, your website and expert texts. Whoever understands which sources the AI reads about their own roofing business can specifically ensure they appear there correctly and completely.

Why the AI suddenly knows about your roofing business

More and more homeowners no longer type their question into Google, but ask ChatGPT, Gemini or the Microsoft Copilot directly: "Who can renew a flat-roof sealing with bitumen in Rosenheim and is certified?" The AI answers with names, assessments and sometimes even a recommendation. The decisive part: it doesn't invent this answer completely, it assembles it from sources available about you on the web. Whoever is not cleanly represented there simply doesn't appear - and loses the job to the business in the next town.

For you as a roofer this means: visibility no longer arises only through ranking positions on Google, but through the question of whether the AI finds enough reliable material about your business. If the material is missing or contradictory, your competition gets named. This is exactly where Generative Engine Optimization comes in, i.e. the targeted preparation for AI systems to understand, classify and recommend your roofing business correctly.

The good news: you don't have to be a tech pro. You just have to know which sources the AI actually reads - and ensure that it gets a clean, complete and consistent picture of your business there. We now go through exactly these sources one by one.

SCORE

Source 1: Your Google Business profile is the foundation

The Google Business profile (formerly Google My Business) is the most important single source. AI systems access this structured data directly or indirectly: company name, address, phone number, opening hours, category and the services you deposit there. If your profile names only "roofer" as the category, but not "sheet-metal work", "flat-roof construction" or "roof-window installation", then the AI lacks this context. It then cannot assign you to a question about zinc cladding or Velux installation.

Enter your services concretely and with technical terms: pitched roof, flat roof, thatched roof, bitumen membrane, roof greening, photovoltaic installation on the roof, storm-damage repair, gutter cleaning. Every service you name there is a signal the AI takes in. Add real photos of completed projects with meaningful descriptions instead of just "Bild1.jpg".

Keep the so-called NAP data - name, address, phone - exactly identical to all other places on the web. If your website says "Dachdeckermeister Huber GmbH" and the Google profile says "Huber Bedachungen", a contradiction arises for the AI. In case of doubt it then decides on a competitor whose data is unambiguous.

Source 2: Business directories and the guild search

AI models also learn from industry portals and directories: the tradesperson search portal of the Chamber of Crafts, the business search of the Central Association of the German Roofing Trade (ZVDH), MyHammer, Das Örtliche, Gelbe Seiten, ProvenExpert. An entry in the official guild or ZVDH database is especially valuable, because AI systems classify such association-backed sources as trustworthy. If you are a guild business, ensure that your entry there is current and complete.

Make sure the same information appears in all directories. Old entries with the mobile number from three years ago or the former workshop address confuse not only customers but also the AI. An afternoon spent going through and correcting the five or six most important portals pays off directly here.

If you have professional distinctions - such as the master title, the specialist-business certification for sealing or a Velux partnership - these belong in the directory profiles. It is exactly by such qualification features that homeowners ask the AI when it comes to a demanding roof.

Source 3: Reviews are literal evidence for the AI

Reviews on Google, ProvenExpert or MyHammer are for an AI model not mere stars, but text that it can evaluate word for word. When a customer writes "The Berger family repaired our storm damage on the tiled roof within two days", the AI links your business with the terms storm damage, tiled roof and fast response. These phrasings from real reviews flow directly into the AI's answers.

So actively ask your satisfied customers for a review and - when it fits - to describe concretely what was done. A review that names the type of roof, the location and the service is significantly more valuable for your AI visibility than a mere "Everything great, gladly again". Five detailed reviews have a stronger effect than fifty meaningless ones.

Also respond to reviews yourself, especially critical ones. Your factual reply is likewise text that the AI reads - and it shows that you care about customers. A composed handling of a complaint about a delayed scaffolding removal says more than any self-description.

Source 4: Your own website as a fact supplier

Your website is the only source you control one hundred percent yourself. AI systems read your texts and draw facts from them. A homepage with the one sentence "We are your roofing professional out of passion" delivers nothing tangible to the AI. A page that clearly says in which radius you work, which roof types you cover, which materials you process and how quickly you're on site in emergencies, by contrast, delivers plenty of usable information.

Create separate subpages for your most important services: one for flat-roof renovation, one for pitched-roof re-covering, one for roof windows, one for photovoltaics on the roof. Write them the way your customers ask. Use real phrasings from customer conversations: "What does a new roof covering for a single-family house cost?" or "How do I recognize whether my flat roof is leaking?". The AI understands such question-answer structures especially well.

Name your region concretely and the places where you are active. For a local roofing business, geographic assignment is decisive. If you work within a 40-kilometer radius around Kempten, then write that out, including the most important neighboring towns. The AI then links your business with exactly these location inquiries.

Source 5: Expert texts, press and mentions from outside

What others write about you often weighs more for the AI than your own self-presentation. An article in the local newspaper about the elaborate renovation of the church roof, a post in the regional portal about your training business, a mention in the blog of a property developer you work with: all of these are independent pieces of evidence that support your profile. AI models weight such external sources as especially credible.

You don't have to hire a PR agency for that. A call to the local editorial office when you complete a special project, a guest article in the newsletter of the local building association or an expert text about the pros and cons of roof greening on your own page can make a start. What matters is that your business name appears in connection with your professional competence.

Cooperations also pay in: if you are on the reference list of a tile manufacturer, a scaffolding builder or a solar installer, that is a mention from outside. Feel free to ask your partners whether they name you as a specialist partner on their website. Every reputable link is an additional signal.

Source 6: Structured data that the AI reads especially cleanly

Besides the texts there is a technical layer that many roofing businesses overlook: structured data in the background of the website, so-called schema markup. With it you deposit for machines unambiguously readable that you are a trade business (LocalBusiness), where you are located, when you're open and which services you offer. For the AI it is like a cleanly filled-out form instead of running text it first has to interpret.

That sounds technical, but in common website systems it is quickly done with an add-on module or by your web provider. Ask your web administrator concretely for "Schema.org LocalBusiness markup" and for FAQ markup for your most frequent customer questions. Both help AI systems capture your content reliably.

Add a real FAQ section with the questions customers ask you constantly: How long does a roof covering last? Do you handle the communication with the insurance in case of storm damage? Do you also work in winter? These answers are worth gold - they are readily adopted directly by the AI and show your competence.

How to proceed concretely: your roadmap

Start with the basics. First check your Google Business profile for completeness and enter all services with technical terms. Then go through the most important directories and align name, address and phone number exactly everywhere. That costs you half a day and eliminates the most common contradictions on which AI systems stumble.

Next take care of reviews and your website. Specifically ask the last satisfied customers for a meaningful review, and build for your three most important services one clearly worded subpage each. Remember: write the way your customers ask, not in advertising language.

Regularly check yourself what the AI outputs. Once a quarter ask ChatGPT or Gemini "Which roofers are there in [your town] and what distinguishes them?" and see whether and how your business is named. This simple check shows you in black and white where you are already visible - and where material is still missing.

Common questions

As a small roofing business, do I really have to do something for AI visibility?

Yes, as a small business in particular. Homeowners increasingly ask the AI instead of Google, especially for special topics like thatched roof, roof greening or storm damage. If your business doesn't appear there cleanly, the AI recommends the competitor. The effort is manageable: a maintained Google profile, consistent directory entries and a few good reviews already get you considerably further.

Which source is most important if I have little time?

The Google Business profile. It is the central data source that almost all AI systems access directly or indirectly. Enter all services there with real technical terms, keep opening hours and contact data current and collect concrete reviews. If you do only one thing, do this. After that follow reviews and your own website as the second and third most important levers.

How does the AI know that I'm specialized in flat roofs, for example?

From the sum of your sources. If you name flat-roof sealing, bitumen membrane and tapered insulation in your Google profile, on a dedicated website subpage, and these terms appear in customer reviews, the AI links your business unambiguously with this specialization. The more concrete and consistent the technical terms are across all sources, the more surely the AI assigns you to the fitting job.

Share