How can an AC contractor tell if AI search is actually sending customers?
An HVAC contractor can measure AI search leads by combining three habits: checking website analytics for referral traffic from AI platforms, asking every new customer how they found the business during intake, and watching for unusual upticks in branded search (people searching the company name directly) or direct site visits after being told about the business by an AI tool. No single signal proves it alone, but together they build a reliable picture.
Signs an AI answer sent a lead
A lead that came from an AI tool like ChatGPT, Gemini, or Perplexity often sounds different from one that came through a Google search ad or a neighbor's referral. These callers tend to describe their problem in more detail upfront, mention that they "asked" something or "looked it up," and sometimes reference a specific recommendation, like a service area or specialty, that matches language from the business's website rather than a review site.
Front-desk staff and dispatchers are usually the first to notice this pattern, even before it shows up in any dashboard. Callers who say something like "I asked an AI assistant for HVAC companies near me" or "it recommended you" are giving a direct, unprompted answer. Training whoever answers the phone to listen for this phrasing, and to jot it down, turns a vague hunch into a running log the owner can review.
Asking new customers how they found you
Every HVAC business already asks some version of "how did you hear about us," but the answer choices often lag behind how people actually search. Adding options like "AI assistant / chatbot," "asked ChatGPT or Gemini," or "AI search summary" to the intake script or CRM (customer relationship management software) dropdown makes this source visible instead of getting lumped into "internet search" or "other."
This works best when it is asked consistently, not just when someone remembers to. Whoever schedules the appointment or answers the phone should ask the same question every time, log the answer the same way every time, and review the tally monthly. Even a rough count kept in a spreadsheet gives an owner something firmer than a guess, and it costs nothing beyond a habit change at the front desk.
Watching branded search and direct traffic
AI tools often name a specific company by name when answering a question like "who repairs air conditioners near me," and when they do, the person hearing that answer frequently turns around and searches the business name directly on Google, or types the web address straight into a browser. This shows up in analytics as branded search volume or direct traffic, and a noticeable rise in either, without a matching rise in paid ads or other campaigns, is a reasonable sign that something outside normal search is putting the name in front of people.
This signal is best read over weeks or months, not single days, because HVAC demand already swings with weather and season. An owner who checks branded search and direct traffic on a regular schedule, alongside the intake answers from new customers, can start to see whether both are moving in the same direction. When they do, that alignment is more convincing than either signal alone.
What a realistic result looks like
A realistic result from AI search visibility is a gradual increase in a specific, trackable category, more customers naming an AI assistant during intake, more branded search activity, more calls that reference being told about the business, rather than a sudden flood of new leads that "prove" AI search overnight. HVAC purchase decisions, especially for larger jobs like system replacement, usually involve a person researching over days or weeks, so the effect tends to build gradually rather than spike.
Owners who expect an immediate, dramatic shift often dismiss AI search as ineffective before the intake data has had time to accumulate. A more useful approach treats this the same way a contractor would treat any new referral source: track it consistently for a few months, compare it to prior months, and judge the trend rather than any single week. A business with steady, if modest, growth in AI-attributed leads over that stretch has a genuine signal, even without a dramatic before-and-after moment.
Checking your own progress without waiting on anyone else's report
An owner does not need a third party to confirm whether this is working. The simplest self-check is a recurring monthly routine: review the "how did you hear about us" tally from intake, look at the website's branded search and direct traffic numbers in analytics, and skim any notes dispatchers or schedulers have logged about callers mentioning an AI assistant. Doing this on the same day each month, like the first Monday, turns it into a habit rather than a one-time audit.
Beyond the numbers, it helps to periodically type the kinds of questions a customer might ask, such as "best AC repair company near me" or "who installs heat pumps in your service area," into ChatGPT, Gemini, or Perplexity and see whether the business gets mentioned, and how it is described. Doing this every few months shows whether the business's presence in AI answers is holding steady, improving, or fading, and it takes only a few minutes. Combined with the intake log and the analytics check, this gives an owner a clear, self-verified view of whether AI search is genuinely sending customers, built from the business's own records rather than anyone else's summary.