AI search for HVAC companies

A homeowner who asks an assistant for a furnace repair company is not using one product. ChatGPT search, Perplexity, and Google's AI features each build an answer from their own sources, and none of them publishes a ranking formula you can optimize against. What you control is the raw material they work from: your site, your listings, and your reviews.

That is why this page is about inputs and not tricks. Our HVAC marketing plan treats AI answers as one surface among several, and the work below is the part of that plan aimed at it.

Google answers the technical question directly. Its documentation says there are "no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." It also says there is "no special schema.org structured data that you need to add." What it does require is that a page be indexed and eligible to show with a snippet.

For a heating and cooling site, that turns into a short checklist: no accidental noindex tag, no snippet restrictions on service pages, content that is in the HTML and not hidden behind a script, and a sitemap that lists the pages you want found. We run that check as part of the AI Overviews work, and the same scan is part of the free audit.

OpenAI runs two crawlers for two purposes. Its documentation says "Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers," while GPTBot crawls content that may be used to train foundation models, and the two are controlled separately. A robots.txt file written years ago, or one a plugin edited, can block one or both without anyone noticing.

Perplexity says that PerplexityBot is "designed to surface and link websites in search results on Perplexity" and "is not used to crawl content for AI foundation models." Google-Extended is a separate control, and Google states that it "does not impact a site's inclusion in Google Search nor is it used as a ranking signal in Google Search."

Whether to allow training crawlers is a business decision, and we make it with the owner. Blocking the search crawlers by accident is simply a lost listing, so our scan reads the robots.txt rules for AI crawlers and flags any that cut you off from search.

An assistant that describes a company can draw on many places: the site, the Business Profile, directories, insurance and supplier pages, and local news. If one directory lists an old phone number, another omits heat pumps, and the site says you serve three counties while the profile says five, the summary can end up repeating the disagreement.

The fix is routine. We run an entity audit to list every place your name appears, then correct the name, address, phone, services, and service area to match one master record. The citation cleanup and consistency pass handles the directories. On the site, structured data can state your services, hours, and service area in a machine-readable form that search engines use to understand a business, and the HVACBusiness type is the specific schema.org type for this trade.

When an assistant recommends a company, it has to say something about it, and customer reviews are one place that language could come from. That link is a reasonable expectation, not something the assistants document, so we treat reviews as worth building for their own sake and for the map results.

For this trade, the useful reviews name the job: a furnace that stopped igniting, a capacitor swap on a hot afternoon, a heat pump installed over two days, a technician who explained the quote. Reviews like those match how people phrase a question, which is more helpful than a string of five-star ratings with no detail.

We build the review request system around asking after every job and replying to each review, and the Business Profile page for HVAC covers where those reviews appear.

HVAC questions fall into two groups. Urgent ones sound like "furnace not heating, who can come today" or "AC blowing warm air, emergency repair near my town." Research ones sound like "is it worth repairing a 15-year-old furnace," "heat pump or gas furnace for a cold climate," and "what does a new air conditioner cost."

Those groups need different pages. The urgent questions are answered by availability, service area, and a phone number in plain HTML. The research questions are answered by pages that give direct, careful answers, which is the HVAC SEO page territory. An owner can ask these question types of the assistants directly, using the towns in the service area, and note which companies are named. Repeating that on a schedule is what AI mention and citation-rate tracking, a GEO deliverable, is for.

The adoption numbers are real. Per BrightLocal's 2026 consumer survey, use of ChatGPT and other generative AI tools for local recommendations rose from 6% last year to 45%.

The visibility numbers are narrower. SOCi's 2026 Local Visibility Index found that "ChatGPT recommends 1.2% of brand locations, compared to an average 35.9% appearance rate" in Google's local 3-Pack. That source measured locations of multi-location brands, not independent contractors, so it describes the gap between the two surfaces for large chains and not for a single-truck shop. The map results still matter, which is the subject of our HVAC SEO page.

We do AI mention and citation-rate tracking by asking the assistants the questions your customers ask, in your service area, and recording which companies they name and which sources they link. The results go into the monthly report next to calls and booked jobs. To see where your company stands today, request the free audit.