Why AI assistants skip your business — and what to do about it

Ask an AI assistant to recommend a plumber in Monument, Colorado. You will get an answer — usually two or three named businesses, sometimes with a sentence about each. What you will not get is a list of every plumber in town.

That is the whole problem in one sentence. Search gave you a ranked list and let you scroll. Assistants give you a shortlist and stop. Being on page one used to mean being considered. Now there is no page one.

Here is what actually determines whether you make the shortlist.

The assistant has to be able to read you

This sounds obvious and is routinely ignored. A large share of local business sites present their most important facts — services, service area, hours, pricing — only as styled text inside a layout. A person reads that fine. A machine extracting structured facts gets much less than you would expect.

Structured data fixes this. It is a block of machine-readable JSON in your page that states, unambiguously:

• what kind of business this is
• where it is and what area it serves
• what services it offers
• how to contact it

This is not a growth hack. It is closer to labeling your inventory. The sites that get cited are disproportionately the ones that made themselves easy to quote.

It has to be able to corroborate you

Assistants are cautious about asserting facts from a single source. If your website says one thing and your Google Business Profile says another and an old directory listing says a third, the safe move for a model is to name a competitor whose details agree with each other everywhere.

This is why the unglamorous work matters so much:

• one consistent business name, in the same form, everywhere
• one address format
• one phone number
• a description that says the same thing in different words, not different things

Consistency is not a style preference here. It is the difference between being confirmable and being a risk.

It has to have something worth quoting

Assistants generate answers by drawing on text. Text that makes a clear, specific, self-contained claim is far more quotable than text that is vague or that only makes sense after three paragraphs of preamble.

Compare:

We pride ourselves on quality service and customer satisfaction, serving the community for years.

with:

We repair and replace water heaters in Monument, Palmer Lake, and Woodmoor, usually same-day, and we have done it since 2011.

The second one can be lifted directly into an answer. The first cannot be lifted into anything, because it does not say anything.

Know which crawler does what before you block one

There is a live debate about whether businesses should let AI companies crawl their content. For a publisher with paid archives, that debate is real.

Before joining it, it is worth knowing that “AI crawler” is not one thing. The names get used interchangeably and they do different jobs. Taking OpenAI’s three, which are documented separately and controlled separately in robots.txt:

• OAI-SearchBot surfaces and links to sites in ChatGPT’s search results. This is the one that affects whether ChatGPT can show you.
• GPTBot collects content that may be used to train OpenAI’s models. Disallowing it is a statement about training, not about search visibility.
• ChatGPT-User is the fetch that happens when a person asks ChatGPT something and it goes to look at a page. It does not crawl on its own.

Google draws a similar line. Google-Extended governs certain Gemini training and grounding uses, and Google states plainly that it does not affect inclusion or ranking in Google Search.

So the honest version of the advice is narrower than the version you usually read. Blocking GPTBot does not remove you from ChatGPT’s search results. Blocking Google-Extended does not remove you from Google. Blocking OAI-SearchBot genuinely does affect whether ChatGPT can surface you.

For a local business trying to be recommended, we would still allow all of them: there is very little to protect and a real cost to guessing wrong about which control does what. But make it a decision. The most common problem we find is not a business that chose to block these — it is a business whose robots.txt blocks them because a plugin or a template did it, and nobody knew.

How to check where you stand

You can do a rough version of this yourself in fifteen minutes:

1. Ask three assistants the question a customer would ask. “Best [your category] in [your town].” Note who gets named. Note whether you do.
2. View your homepage source and search for application/ld+json. If there is nothing, you have no structured data.
3. Check your robots.txt at yoursite.com/robots.txt for AI crawler blocks you did not intend.
4. Search your business name and compare the top five results. Do the name, address, and phone match exactly? Exactly, not approximately.

If that turns up problems and you would rather have someone else work through it, the free Visibility Snapshot does exactly this and costs nothing.