A homeowner in Clifton asks ChatGPT who to call for a burst pipe. ChatGPT names three plumbers. If one of them is not your shop, the reason is checkable in 20 minutes with a phone or a laptop. The audit below needs no paid tools and no code.
How ChatGPT picks local businesses runs a five-step pipeline: query expansion, entity resolution, trust scoring, content evaluation, and synthesis. The audit below checks the inputs to that pipeline, one at a time, from the perspective of a trades owner who wants to know what the AI sees.
Step 1: Search the service, not the shop (5 minutes)
Open ChatGPT. Type a query a homeowner would use, not your business name. Try these:
- "Who is the best plumber in [your city]?"
- "Who does emergency AC repair in [your county]?"
- "Find me a licensed electrician for a panel upgrade in [your town]"
Write down every business ChatGPT names. If your shop is on the list, note where it ranks. If it is not, that is the baseline.
Repeat the same query from a second zip code inside your service area. AI answers change by location. A shop that appears in one zip and disappears in another has a geographic anchor problem, not a review problem.
Run the same prompts on Gemini and Perplexity. Each model weighs signals differently. Gemini favors websites. ChatGPT leans on listings. Perplexity spreads across directories. A shop missing from all three has a structural gap. A shop missing from one has a model-specific gap.
Step 2: Check the Google Business Profile (3 minutes)
Open Google Maps and search your business name. Confirm three things:
- The profile is claimed and verified (look for the "Own this business?" button, which means it is unclaimed)
- The primary category is the specific trade, not "Contractor" or "Home services"
- The service description names the city and the trade
The Google Business Profile is the AI strategy because it is the listing every AI model checks first. Yext's July 2026 research found that 42% of all AI citations come from business listings, and the Google listing is the most-cited one.
If the category says "Plumbing repair" instead of "Home services," the AI can match it to a plumbing query. If it says "Home services," the AI cannot match it to anything specific enough to cite.
Step 3: Check NAP consistency across directories (4 minutes)
Search the business name on Yelp, Bing Places, Apple Maps, and Angi. For each one, confirm the name, address, and phone number match the Google listing exactly.
Common mismatches that break entity resolution:
- Yelp says "Smith Heating & Cooling LLC" but Google says "Smith HVAC"
- Bing Places has an old phone number from before the switch to VoIP
- Angi lists the shop under "Heating" instead of "HVAC"
- Apple Maps shows a different service area than the Google profile
NAP consistency is how AI engines confirm that two listings point to the same business. A mismatched name or phone breaks the link. The model sees two partial businesses instead of one complete one, and neither has enough signal to recommend.
Step 4: Check the website for structured data (3 minutes)
Open the business website. Right-click and select View Page Source. Search for "LocalBusiness" or "schema.org" in the HTML.
If LocalBusiness schema is present, confirm it includes: business name, address, phone, hours, service area, and aggregate rating. If any of those are missing, the schema is incomplete.
If no schema is present, the website is prose to the AI. Gemini pulls 52.1% of its citations from first-party websites, per Yext's July 2026 study. A website without structured data gives Gemini nothing to extract. Brand-managed sources account for 86% of AI citations, and the website is the largest single slice.
Step 5: Check the star rating floor (2 minutes)
Count the Google review average. SOCi's 2026 Local Visibility Index found that ChatGPT applies a 4.3-star average as a floor before it will recommend a business. Star ratings are a gatekeeper for AI visibility, not just for Maps.
If the average sits below 4.3, the fix is review velocity. Ask the last ten customers for a review that names the service and the city. "Fixed the AC in our Clifton home same-day" tells the model the trade, the town, and the quality. "Great service" builds a number but not an entity.
Step 6: Check whether the AI can confirm the service area (3 minutes)
Go back to ChatGPT. Type: "What services does [your business name] offer in [your city]?"
If the AI cannot answer, or gives a vague response, the service area is not specific enough on the website or the Google profile. The fix is to add city-level service pages. One page per town in the service area, each titled with the trade and the city: "Plumbing Repair in Paterson, NJ" or "HVAC Installation in Wayne, NJ."
A single page that says "serving North Jersey" gives the AI nothing to pin to a grid point. Paterson, Wayne, and Passaic are pinable. "North Jersey" is not.
The full 20-minute checklist
- Search the service and city on ChatGPT, Gemini, and Perplexity from two zip codes
- Confirm the Google Business Profile is claimed, categorized, and city-specific
- Match name, address, and phone across Yelp, Bing, Apple, and Angi
- Confirm LocalBusiness schema is on the website with all fields
- Check the star rating against the 4.3 floor
- Ask ChatGPT what services the business offers in the city
If all six pass and the shop still does not appear, the gap is third-party mentions. The AI cross-checks what the business says about itself against what other sources say. A shop with no mentions on Reddit, local news, or industry directories has a thin trust layer. That is a longer fix, but it starts with getting listed on the directories the AI already reads.
Run a free AI visibility audit to check all 12 signals across ChatGPT, Gemini, and Perplexity in one pass. The audit covers the six steps above plus six more, including schema validation, review sentiment, and citation gap analysis.
FAQ
How do I check if ChatGPT recommends my trades business?
Open ChatGPT and type a service query with your city, not your business name. Try "who is the best plumber in [your city]" or "who does emergency AC repair in [your county]." Write down every business ChatGPT names. If your shop is not on the list, the AI cannot find it for that query. Repeat on Gemini and Perplexity.
What should I search for when auditing my trades business on AI?
Search the service a homeowner would type, not your shop name. Use specific trade queries like "emergency plumber in Paterson NJ" or "licensed electrician for panel upgrade in Bergen County." Search from two different zip codes inside your service area. The results change by location.
Why does ChatGPT not recommend my business even with good reviews?
Reviews are one signal. ChatGPT also needs a claimed and complete Google Business Profile, consistent name-address-phone across directories, and structured data on the website. Yext found 86% of AI citations come from websites and listings businesses already control. If any of those are missing or inconsistent, the AI skips the shop regardless of review count.
How often should I audit my trades business on AI search?
Run a manual check once a month. AI recommendations have roughly 85% volatility, meaning the businesses AI names can shift between queries. A monthly check catches drops before they become a pattern. FirstCall's free audit runs all 12 signals across three AI platforms in one pass.
Can I audit my business on ChatGPT for free?
Yes. The free tier of ChatGPT, Gemini, and Perplexity all answer local service queries. Open each one, type the service and city a homeowner would use, and write down who gets named. The full FirstCall audit at firstcallvisibility.com runs 12 signals across all three platforms at no cost.
Run your free AI visibility audit at firstcallvisibility.com.