A moving company with 181 Google reviews and a 4.8-star rating scored zero in Google AI Overviews for its own core service. Foxxr Digital Marketing published that result in July 2026 after tracking 50 home service businesses across 12 states and 15 trades. Seven of those shops never appeared. Three more scored under 10.

That is the gap most trades owners miss. Google AI Overviews for contractors are a different surface from the local pack. A shop can sit in Maps, collect five-star reviews, and still get skipped when a homeowner asks Google who to call.

What Foxxr actually measured

Foxxr ran one natural-language prompt per shop through Local Falcon's geospatial grid and scored how often Google's AI Overview named that shop across its service area. The metric is Share of AI Voice, 0 to 100. A 100 means the shop showed up at every tracked grid point. A 0 means it never did.

The median score was 71, which looks fine until the tails. Nine shops scored a perfect 100. Those nine, 18 percent of the sample, accounted for 28.5 percent of all Share of AI Voice points in the study. The other 41 averaged 55.1. Thirteen shops, more than one in four, scored below 40. Foxxr treated 40 as the line where AI recommendation starts to look like a real lead source.

Foxxr sells AI visibility work, and almost all 50 shops were clients. The pattern is still one a trades owner can check on a Saturday. Search the service, not the shop name. Then search again from a different zip code inside the same county.

The Clifton HVAC that looks strong on paper

Take a Clifton, New Jersey HVAC shop with 4.7 stars, 90 reviews, and a Google Business Profile set as a service-area business so the house address stays hidden. The website hero says "HVAC services across North Jersey." The service page is titled "Air Conditioning Repair."

A homeowner in Wayne types "who repairs AC in Wayne NJ" into Google. The Overview names two competitors. Foxxr's 100-score example was a Wall Township mold shop whose top page was titled "Mold Removal and Remediation Services in Wall Township, NJ." That page named the job, the town, and the state. Google could match it to a GBP, reviews, and third-party mentions that all said the same thing.

The Clifton shop has reputation. It does not have a confirmable place. "North Jersey" is not a city Google's AI can pin to a grid point. Wayne, Paterson, and Passaic are.

Three reasons a shop scores zero

Foxxr grouped the seven zeros into three patterns. Some shops hit more than one.

Entity confusion. One shop lost to a competitor with a near-identical name in the same market. The AI could not tell which entity to recommend and picked the cleaner signal set. Trades businesses that added "Heating & Cooling" on Angi, dropped "LLC" on Yelp, and kept a DBA on the truck run into the same mess. Every listing has to point at one name.

Missing geographic anchor. Three shops had no findable Google Business Profile. Two had none. One had a profile that was not in Google's local entity graph. AI Overviews lean on that graph for service queries. Without a claimed, verified, categorized profile, the model defaults to shops that have one. The Google Business Profile is the AI strategy for a reason.

Thin trust layer. Three shops had a profile that was too empty for the model to trust. Zero reviews. Incomplete hours. A pretty website with nothing for a third party to confirm. The AI cross-checks what a shop says about itself against what other sources say. Silence is not a source.

The moving company sat outside those three. Reviews were strong. The website was not. Service pages were written to impress visitors instead of stating the job, the town, and the customer. The AI knew people liked the company. It did not know what to recommend it for.

Star ratings as a gatekeeper still matter. SOCi's 2026 Local Visibility Index found ChatGPT recommendations clustering around 4.3 stars. Foxxr's mover cleared that bar and still scored zero on Google's Overview grid. Stars are a floor. They are not the whole building.

Coverage across the service area, not one zip

A shop that appears for a homeowner on the same block as the shop and disappears ten miles away has a Share of AI Voice problem, even if the owner tested the prompt once from the office. Foxxr's full-grid point is the one trades owners should steal.

Passaic County HVAC work does not happen in one town. A Saturday check for a North Jersey shop looks like this:

Search "emergency AC repair in Clifton NJ" from a Clifton zip.

Search "who installs heat pumps in Wayne NJ" from a Wayne zip.

Search "licensed HVAC in Passaic County" from a Paterson zip.

If the Overview names the shop in one of those and a competitor in the other two, the site is not giving Google enough town-level pages to confirm. A single "North Jersey" page will not cover three cities.

All 50 tracked businesses already had blogs. Fourteen percent still scored zero. Volume was not the split. Confirmable content was. A post about signs of a bad capacitor is informational. A post about a 1998 gas furnace in a Wayne Cape that lost cooling in July is locatable. Reviews work the same way. "Great job" builds a rating. "Replaced the condenser in a Clifton ranch, same-day" tells the model the trade and the town.

What to fix this week

Open Google and search the service a homeowner would type, not the shop name. Do it from two zip codes. Write down who the Overview names.

Open the three busiest service pages. If a page does not name a city, a job, and a problem in the title or first paragraph, rewrite that page first. "AC repair in Clifton, NJ" beats "Quality HVAC services."

Open Maps and search the shop name. If a similarly named competitor surfaces first, fix the GBP name, title tags, and directory listings so they match.

Read the last ten reviews. If none mention a service and a city, ask the next customer to include both in one sentence.

None of that requires a franchise or a new website. It is the punch list FirstCall runs in a free AI visibility audit. Why ChatGPT does not recommend a business is a related check. Google's Overview grid is the one Foxxr measured, and the one a homeowner sees without opening ChatGPT.

FAQ

Can a contractor with 4.8 stars still be missing from Google AI Overviews?

Yes. Foxxr's July 2026 study tracked a moving company with 181 Google reviews and a 4.8-star rating that scored zero Share of AI Voice for its core service. Stars help. Google's AI still needs a confirmable service, a named town, and a claimed Google Business Profile.

What is Share of AI Voice for a trades business?

Share of AI Voice is Foxxr's 0-to-100 score for how often a shop appears in Google AI Overviews across a local grid, not just at the shop's own address. A 100 means the shop showed up at every tracked grid point for its core service query. Below 40, the shop is missing across most of its own service area.

Why does an AI Overview change from one neighborhood to the next?

Google's AI Overviews pull different answers depending on where the searcher is sitting. A shop can show up in one zip code and disappear ten miles away. Service pages that only say North Jersey give the model nothing to pin to Clifton, Wayne, or Paterson.

What should a trades shop check first if it is missing from AI Overviews?

Search the service, not the shop name. Type who does AC repair in [city] from two zip codes inside the service area. Then confirm the Google Business Profile is claimed, verified, and categorized, and that the main service page names the city and the job.

Do blogs help a contractor show up in Google AI Overviews?

Blogs help when they name a service, a town, and a problem the model can confirm against the Google Business Profile and reviews. All 50 shops in Foxxr's sample had blogs, and 14 percent still scored zero. Generic how-to posts without a city attached did not get those shops recommended.

Run a free AI visibility audit at firstcallvisibility.com.