You ask ChatGPT who the good plumbers in your town are. It names four companies. You are not one of them, and you have been doing this for eleven years.
Or worse, it does mention you, and half of what it says is wrong.
Here’s the short version of why. These systems don’t rank businesses the way Google does. They assemble an answer from what the web appears to agree on, and if the web can’t confirm you exist as a specific, consistent business, you get left out of the list. Not penalized. Just skipped.
If you want to find out what it currently says about you, I wrote that up separately. This one is about why you’re missing, and what to do about it.
The thing most people get wrong about this
Business owners assume this is a content problem. Write more blog posts, get mentioned more, climb the rankings.
It’s usually not. It’s an identity problem.
An AI answering “who should I call in Wesley Chapel” needs to be confident that a particular business exists, does that specific thing, and operates in that specific place. It builds that confidence by finding the same facts in several independent places and seeing them agree. Your website, your Google listing, your LinkedIn, your directory listings, review sites.
When those sources disagree, or when most of them simply don’t exist, the model has nothing solid to hold. So it names the four businesses it is sure about.
Check these, in this order
One. Make sure you’re not blocking them. Go to yoursite.com/robots.txt and look for GPTBot, ClaudeBot, PerplexityBot, Google-Extended or CCBot. If any of those are disallowed, you’ve opted out of the thing you’re trying to appear in. A surprising number of sites did this in 2023 without understanding the trade and never revisited it.
Two. Check your name is actually your name. This sounds idiotic. It is not. More on this below, because I had this exact problem and did not notice for months.
Three. Claim and fill your Google Business Profile properly. For a local business this is the single strongest confirmation that you exist, you’re real, and you’re where you say you are. Half-finished profiles are the most common cause of local invisibility I see.
Four. Get your details identical everywhere. Name, address, phone. Identical, not approximately the same. “Suite 4” on one and “Ste. 4” on another is the kind of thing that weakens a match. If you changed your number two years ago, the old one is probably still sitting in six directories.
Five. Make sure the answer exists in plain words on your site. If someone asks what you charge and nobody in your industry publishes a number, the model has nothing to quote. The page that answers the question directly is the page that gets used. This is also why reviews matter more than people think. They’re other people describing your business in their own words, which is exactly the independent confirmation these systems look for.
What I found on my own site
I do this for a living and I had three of these problems.
My site was publishing my name, in the structured data every page carries, as JMIKEOLIVER. One word, all capitals. Not a person’s name, not a company name, nothing anybody would ever search. It came from a WordPress display name I set once in about four seconds and never looked at again. Every page of the site had been quietly telling machines that for months.
Second, the field that lists your other profiles was completely empty. That field is how you say “this website and this LinkedIn are the same entity.” With nothing in it, my site was an island. No corroboration anywhere.
Third, there was no business information in the structured data at all. No phone number, no email, no service area. A page can say “Tampa digital strategist” in the headline and still carry nothing a machine can read as a location.
Fixing it took about twenty minutes. The display name, one profile link, a business block with the phone number and the towns I actually serve, and a service declaration on each service page. None of it changed a word of what visitors see.
The one thing I had right by accident was robots.txt. Nothing was blocked. If it had been, everything else would’ve been wasted effort.
When it says something wrong about you
Different problem, and more annoying, because you can’t edit the answer.
Wrong information usually traces back to a real source. An old address on a directory nobody has updated since 2019. A news article from when you had a different owner. A competitor’s page that mentions you in passing. The model isn’t making it up, it’s repeating something it found.
So the fix is the same shape: find the stale source and correct it where it lives. Update the directory. Fix the old listing. Publish the correct version clearly on your own site, since your own site carries real weight on questions about your own business.
What doesn’t work is arguing with the chatbot. There’s no form to submit, no appeal, and the answer varies from one person to the next anyway.
What you can’t control
Plenty. These models update on their own schedule, two people asking the same question get different answers, and a small local business will never be as confidently known as a national chain.
So don’t measure this in rankings. There aren’t any. Measure it by asking the question yourself every month or so and watching whether you start turning up, and whether what it says is closer to true than it was.
And keep it in proportion. For most local businesses this is still a smaller channel than Google and much smaller than word of mouth. It’s growing quickly enough to be worth an afternoon. It isn’t worth abandoning what already works.
The honest summary
If you’re invisible, the problem is usually that you haven’t given these systems enough to be sure about. Not enough corroboration, not enough consistency, and often a basic detail published wrong in a place you’ve never looked.
Start with robots.txt, then your Google Business Profile, then make your details identical everywhere, then check what your own site is actually publishing about you. That order is roughly cheapest and fastest first.
If you want somebody to look at what your site is currently telling machines about your business, send me the address and I’ll tell you what I find. It’s usually something small and dumb, and it’s usually been that way a while.