Research & insights

The HVAC Company You Found Online Might Not Be an HVAC Company

Dana Lampert · November 22, 2026 · 6 min read · AI Visibility

In February 2024, a Google search for an HVAC contractor near Buffalo Grove, Illinois, returned a business called Levine Heating and Cooling. The listing presented it as a local HVAC contractor that had been operating for more than five years.

According to a federal complaint filed in May 2026, the company didn't exist.

Levine Heating and Cooling was allegedly one of more than 15,000 fabricated Google business profiles created by a Chicago-area home-repair operation. The government says the profiles used invented company names, local addresses and phone numbers, and favorable reviews to persuade homeowners they were contacting established local providers. Calls were routed to representatives elsewhere, and jobs were assigned to third-party technicians.

These are allegations in a pending civil case, not established findings of liability. But they expose a problem that gets overlooked in discussions of search rankings and AI recommendations.

Before deciding whether a contractor is good, we should establish whether the contractor is actually the business we think we're hiring.

A convincing local listing doesn't require a local operation

The Department of Justice's description of the case is remarkable for how ordinary the consumer experience allegedly appeared to be.

A homeowner needed a plumber, electrician, heating contractor, or garage-door repair company. They searched online and found a business with a plausible name, a nearby address, a local phone number, and positive reviews. They called, expecting to reach that business.

According to the complaint, many addresses were fabricated or belonged to unrelated businesses. Calls went to overseas centers, where representatives sometimes collected fees without knowing whether a technician was available. The government alleges that some dispatched technicians were unqualified and others never arrived.

The important point isn't that every dispatched technician was fictitious. Some actual people performed jobs. It's that consumers allegedly believed they had chosen one local company when they were dealing with an entirely different operation.

A company can profit from acquiring calls without maintaining the technicians, equipment, or local presence a homeowner assumes it has.

That doesn't describe every lead-generation business. Plenty of legitimate marketplaces and referral services openly connect customers with independent providers. Contractors can also lawfully use subcontractors, subject to applicable licensing and other requirements.

The distinction is whether the intermediary represents itself honestly and whether the customer knows who is responsible for the work.

A directory telling you it will introduce you to a contractor is one thing. A fabricated contractor listing pretending to be the company that will fix your furnace is another.

This isn't just an obscure loophole

Google says it removed more than 13 million fake Business Profiles during 2025, across business categories worldwide.

That figure does not mean 13 million fake HVAC companies reached consumers. Many fraudulent listings are intercepted or removed, and Google doesn't report an HVAC-specific prevalence rate in that disclosure.

Google also has safeguards. Its Business Profile policies exclude lead-generation companies, and Local Services Ads applies license, insurance, and background screening where required.

The removals show Google is actively fighting this abuse.

Yet the federal case illustrates how persistent the incentive to impersonate a local business can be. It also reveals a weakness in how consumers often evaluate companies: a search result can establish that a business is visible without establishing who is actually behind it.

For the homeowner, the question remains surprisingly basic: who am I actually dealing with?

AI could make the identity problem more consequential

So far, much of this problem has played out through Google Search and Maps. But the interface through which customers find a contractor is changing.

AI assistants now assemble local-business shortlists and can sometimes initiate quote requests or appointments. Google added home-service booking links to AI Mode in November.

There's no established evidence that these assistants systematically recommend fake contractors.

But consider what happens when an AI assistant encounters a convincing business listing without enough evidence to distinguish a real operator from a marketing intermediary.

If the assistant repeats the name, address, rating, and services as though they describe a single accountable company, it can carry the original confusion into a more confident-looking recommendation.

That's a different problem from an AI model getting a contractor's pricing or specialties wrong. The identity of the company itself may be misrepresented.

For an agent acting on someone's behalf, I think verification needs to start even earlier than comparing credentials and customer reviews.

Who is the provider, and who will actually perform the work?

A real operating history is harder to manufacture than a nice listing

There are several pieces of evidence I'd want to see before an AI agent treats a home-service listing as an established operating business.

Start with identity. What legal entity stands behind the name, invoice, warranty, and contract? A trade name is perfectly legitimate, but it should be traceable to whoever is responsible for the work.

Then licensing. When required, is the license active, does it cover the relevant trade, and does it belong to the responsible business or professional? Rules vary by jurisdiction, and printing a license number on a website isn't the same as verifying its standing with the appropriate authority.

Finally, operating activity. Does the company have a documented history of completed work, with identifiable job types, customers, dates, and geography, rather than only marketing claims?

At TrueSignal, this is part of the reason we're building TrustRecord. For businesses that connect their operating systems, we can compute claims such as completed jobs, customers served, service categories, and geographic operating history from source records. Relevant license claims can also be checked against authoritative registries, with the verification method and date published alongside the claim.

Take Sunshine Services, an HVAC, plumbing, and electrical company based in Gainesville, Florida. Its TrustRecord documents actual operating activity computed from its ServiceTitan records, including completed jobs, customers served, the types of work performed, and where that work took place.

The record also includes licenses, insurance, certifications, and operating policies, clearly distinguishing independently computed evidence from information attested to by the company.

That provides a much more substantive picture of an operating business than a Google listing, a polished website, and a collection of five-star reviews.

But operating records require interpretation, too. A completed-job count doesn't prove workmanship or establish which technician physically performed each job. A lead broker could also maintain records of work performed by subcontractors. And a license covering one trade doesn't establish qualifications for another.

Those qualifications are precisely why the evidence needs to be separated, defined, and labeled.

A TrustRecord should provide evidence, not just another badge

We don't need another online badge saying a company is "real" without specifying what was checked.

A free TrustRecord can organize public identity information and facts confirmed by the business, but owner-attested statements should remain labeled as such. A Connected TrustRecord can add operating claims computed from authorized business systems, with their sources and freshness documented. Individual credentials should be labeled verified only when the appropriate registry or supporting evidence has actually been checked.

These distinctions matter.

A license verification establishes what an authoritative registry reported at the time of the check. It doesn't guarantee the business is qualified for every type of work it advertises.

A history of completed jobs provides evidence of operating activity. It doesn't independently establish whether those jobs were performed well, whether customers were satisfied, or who will show up tomorrow.

And a business's statement that it is insured is not equivalent to independently verifying current coverage.

Still, consumers and AI systems should be able to distinguish a company with documented operating history from a listing whose primary purpose may be acquiring phone calls.

The same principle applies beyond home services. A legal referral website isn't necessarily a law firm. A healthcare directory isn't the medical practice that will treat a patient. Intermediaries can be useful, but they should be identifiable as intermediaries, and the underlying provider should be evaluated on its own evidence.

The objective isn't to eliminate referral businesses or impose a single definition of what a legitimate service company looks like.

It's to make clear who is responsible for delivering the service and what evidence supports the company's claims.

The first question should be whether the business actually does the work

The FTC advises homeowners to check licenses, insurance, and references, and obtain written estimates. That remains sensible regardless of how the business was discovered.

But as AI increasingly conducts the initial research, we should expect it to help with more of that basic diligence before presenting a shortlist.

Agents should distinguish operators from referral services, identify who is responsible for the job, and flag unverified facts rather than substituting a polished listing for proof.

A business with legitimate operations should also have an opportunity to document those operations in a way that's accessible beyond its own marketing materials. That means business identity, appropriate credentials, and meaningful operating evidence, each with clear sources and limitations.

The industry spends a great deal of time measuring which companies rank highest, get mentioned most often, or appear in the most AI recommendations.

The federal fake-listing case suggests a more basic question belongs ahead of all those measurements.

Before an AI agent decides which contractor to recommend, it should establish which company is actually going to show up and do the work.

Your business has verified data that's hidden.
A TrustRecord makes your operating history readable by every AI system making recommendations.

Continue reading

AI Is Helping Families Choose Senior Living. The Tour Still Matters.Your Customer Already Asked ChatGPT. Here's What They Got Wrong.When Will We Actually Let AI Choose the Professional?

Create your TrustRecord

Tell us about your business. Takes 5 minutes.

We'll review your business and reach out within one business day.