AI visibility for US local services: publish the license number a state board can confirm

Licensed and insured is a claim. A license number with the board that issued it is a fact, and that difference decides whether an assistant will name an American local service business. When somebody asks ChatGPT or Gemini for a plumber, an electrician, an HVAC company or a roofer in a named city, the answer is assembled from pages that carry checkable details, and the single cheapest checkable detail a licensed trade owns is its state license number. In our benchmark run on 15 October 2026, ChatGPT and Gemini each answered 102 buying questions in the United States, which is 204 US readings, and cited our own domain zero times on either engine. Across the full run of 597 readings, 85.6 per cent of 10,652 citations went to somebody's own website, and review directories took 0.8 per cent.

What we measured, city by city

We then measured cities directly, because a market figure is not a local figure. Ten cities, six buyer questions each, both engines, 120 readings on 15 and 16 October 2026, all of them usable. Across those 120 readings we were not cited once. The four American cities in that set looked like this.

City	Readings	Returned sources	Cited us	Most cited local sources
New York	12	9	0	localseo.nyc 3, thenycseo.com 3
Los Angeles	12	11	0	agencies.semrush.com 4, brightlocal.com 3
Chicago	12	12	0	chicagoseoagency.co 3, paramdigitalmarketing.com 3
Houston	12	11	0	houdigitalmarketing.com 3, thedigitalelevator.com 2

The pattern across all ten cities is the encouraging part. Every city has its own winners and they are small, city specific domains rather than national brands. That is the whole argument for writing like a local business instead of like a national one, and it is reachable this quarter. These are the cities we measured, so there is no figure here for a city that was not in the set, and for one of those the market number is what applies.

Two boundaries on all of the above. Both runs cover two engines, ChatGPT and Gemini, on the stated dates. The product itself covers five surfaces, named: ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews. Both runs on this page are two of those five, so nothing here describes the other three. And the questions were buyer questions asked market by market and city by city, not one study per trade, so there is no plumbing figure and no roofing figure in either run. The method and the full citation table are in what AI assistants actually cite across 597 readings of ChatGPT and Gemini in three markets, with the American slice in the US cut of those 204 readings.

The license number, and the rest of the checkable set

Put these in text on your site, not in a badge image, and name the body that issued each one.

• Your state contractor or trade license: the number, the classification, the state board that issued it, and the status. Most state boards publish a license lookup, so a number is a fact a stranger can confirm in under a minute.
• Your bond and your insurance: the surety, the bond amount as you are permitted to state it, and a line saying a certificate of insurance is available on request with the carrier named.
• Federal certifications where the trade needs them, such as EPA Section 608 certification for technicians handling refrigerants, or lead safe certification for renovation work in older housing.
• Manufacturer and trade credentials, named exactly: the brands you are an authorised dealer or installer for, and the trade certifications your technicians hold.

The reason this works is the same reason the engines asked for it. 102 of 102 build spec conversations in our run demanded a date and a last verified line on any claim, and 80 of 102 asked for the number behind a number. A license with a board, a status and a date satisfies all of that, and it is already in your filing cabinet.

Price, area and hours, written as facts

Every one of the 102 build specs also asked for pricing in the reader's own currency. For a trade that quotes per job, that means publishing the diagnostic or call out fee in dollars, the typical range for your three most common jobs, what moves a quote up or down, and whether evenings and weekends are charged differently. A page with a dollar figure on it gets quoted for a cost question, and a page that says call for a free estimate hands that question to a competitor who published one. What a package of local work usually costs per location, line by line, is set out in local SEO services and the line items it has to name.

Then write your service area as a plain list of the towns, neighborhoods or ZIP codes you genuinely cover, and your hours including holidays and after hours availability. The identity and listings work underneath all of this, keeping one set of trading facts everywhere, is in AI visibility for a local business. This page is the American layer on top: the license, the bond, the fee and the area.

The part of the answer you do not own

Maps data, directory entries and review platforms carry your details whether you maintain them or not, and a stale entry is quoted as confidently as a current one. The useful discipline is to keep the handful that matter accurate and stop there. Review directories were 0.8 per cent of the 10,652 citations in the run, so a listings campaign is not the lever it is sold as. Correct the listings that are wrong, make the license number and the fee live on your own pages, and let that be the week's work.

Running the check from where your customers are

1. Ask in the way a customer does, with the city or neighborhood in the question and your business named nowhere in it.
2. Run five shapes: the emergency job, the planned replacement, the cost question, the who is licensed question, and a shortlist question for that city.
3. Ask each one twice in ChatGPT and twice in Gemini, from a signed out or temporary chat.
4. Save who was named and every source shown, then fix the two facts the engines went looking for and did not find on your site.

Stated plainly, the fix is what we sell. AI Knows Us puts the license, the fee and the service area into pages for a trade and then watches what the assistants do with them, city by city. 64 of the 102 build specs asked for that declaration by name, and the figures above include our own: 3 citations in 597 readings, none in the United States, and zero across all ten cities.

Questions about AI visibility for a licensed trade in the United States

Is publishing our license number a security risk?

It is public information, published by the state board that issued it and searchable by anybody. Putting it on your own site simply means the buyer who wants to check can do it without leaving your page.

We are a franchise. Whose facts should be on the page?

The local entity's, because that is what the question is about. The license held by your location, the owner, the service area and the local phone number, with the brand named alongside rather than instead of them.

Why is a small competitor named when we are larger?

Because their pages are more specific to the city being asked about. In the ten cities we measured on 15 and 16 October 2026, every city was won by small, city specific domains, which is a position earned by writing about one place properly.

Do we need a page for every town we cover?

One page for each service and town you really serve, carrying the local detail only you have: the response time, the typical job in that area, the price range and the permit you pull there. A template with the town name swapped in has nothing specific to quote.

The first number to put on your own site

If you change one thing this week, put your license number, the board that issued it and its status into the text of your home page and every service page, next to the city you serve. It takes ten minutes, it is the one fact in this trade that a stranger can verify without talking to you, and it is the kind of detail that separated the pages getting cited from the pages getting ignored across 597 readings. Then publish your call out fee in dollars, and you have given both a customer and an assistant a reason to pick you out of a list.