Reason 1. AI crawlers cannot get in
The most common one, and the quickest to check. Type yourdomain.com/robots.txt into your browser. If you see Disallow: / under User-agent: * or next to any AI crawler, the case is closed and no other work matters.
The block is almost never deliberate. Usually it came from a plugin, the agency that built the site, or the WordPress setting called "Discourage search engines from indexing this site". Details in the post on AI crawlers in robots.txt.
Reason 2. Your details disagree across services
Hours in Google that differ from Apple Maps. A phone number with the country code in one place and without it in another. An old address in some directory. To a model that reads as unreliable information, and models are trained to be careful with unreliable information about places.
The result is that the model names a neighbour whose details are dull and consistent, instead of you.
Reason 3. The category is too broad
Your profile says "Contractor" and the customer asks for a roofer. The primary category decides whether you are considered at all for a given question. Too broad a category is the cheapest mistake on this list to fix.
Reason 4. There is nothing to read about you
The site is one photo of the menu, a form and a map. The profile has a name, an address and a phone number. The reviews are stars with no text. The model arrives, reads, and finds not one sentence it could quote.
This is the most common case among businesses that do everything right and never wrote any of it down.
Reason 5. The reviews are old
A hundred reviews from four years ago are weak evidence of anything. The place may have changed owner, menu and hours. A dozen reviews from the past year weigh more than two hundred from five years back.
Reason 6. Your business does not match how people ask
The reason few people consider, and sometimes the decisive one. A metal components manufacturer in a village will not get named, because nobody asks an assistant for a components manufacturer in that village. Those customers search differently.
AI visibility makes sense where the customer searches close to home and phrases the question with a place name: food, hair, dental, repairs, florists, hotels, gyms. If your trade does not work that way, we will say so rather than sell you something that changes nothing.
How to tell which reason is yours
- Open
yourdomain.com/robots.txt. Reason 1 is confirmed or ruled out in half a minute. - Search your business name and compare the details across your Google profile, Apple Maps and Facebook.
- Look at the primary category on your profile.
- Ask an assistant about your trade and town five times from a private window. Note who came up.
- Now ask about your business by name. If the model can describe it, the data exists, it just is not trusted enough to volunteer.
Step five is a useful differential test. A model that knows you when asked directly but skips you on a general question has a confidence problem, not a data problem.
The systematic version of this check is the free scan: we ask ChatGPT three of your customers’ questions, read the answers and show them in full with timestamps, along with the causes we found on your side. The result appears straight away, on screen and by email.