When a customer asks ChatGPT for the best option in your category and your competitors are named instead of you, there is rarely one setting to flip. OpenAI's published documentation does not list what decides which businesses get named, so the honest approach is to test, read what ChatGPT cites, and compare that with what exists about you. If you want the short diagnosis first, the not showing up in AI answers page covers the common causes, and our GEO service page describes how we measure and fix them.
Why aren't we appearing in ChatGPT search results?
There are two broad possibilities: ChatGPT cannot reach your pages, or it can reach them and still names someone else. OpenAI documents the first in detail and says very little about the second, so the first job is finding out which one you have.
Here is what OpenAI's help center does say. ChatGPT may search the web automatically when a question would benefit from current information, so not every answer comes from a fresh search. ChatGPT search sometimes partners with other search providers, and when it does, it typically rewrites your question into one or more targeted queries for them. OpenAI also states that ChatGPT ranks search results using multiple factors intended to help users find relevant, reliable information, and that placement is not guaranteed.
That help article does not list the factors, and none of the OpenAI pages cited here says whether reviews, directories, comparison pages or third-party mentions raise or lower your chances of being named. Anyone who tells you exactly how much each one counts is going beyond what OpenAI has published, so everything below is a method for gathering evidence, not a formula.
Rule out crawler access first, because it is documented
Check that OpenAI's search crawler can reach your site first, because this is the part OpenAI documents. Its help center says to allow OAI-SearchBot to crawl the site and to confirm that your host or content delivery network allows traffic from OpenAI's published searchbot IP addresses. OpenAI's crawler documentation adds that sites that opt out of OAI-SearchBot will not be shown in ChatGPT search answers, though they can still appear as navigational links.
Two details save time. Per the same documentation, changes to robots.txt can take about 24 hours to take effect for search results, and GPTBot is a separate setting that concerns training, so you can allow search and disallow training. Passing this check makes you eligible. It does not explain why a competitor is named instead.
Run the test the way a customer would
Write the questions a real buyer would ask, run each one more than once, and record every business named and every source cited. One answer is one sample, and OpenAI says search results and citations can be incomplete, outdated or incorrect, so do not draw conclusions from a single run. Ten questions is a workable start for one person.
Control the things OpenAI says influence the query. ChatGPT may use an approximate location from your IP address and, if memory is enabled, saved memories when it rewrites a search, and a VPN can change the location it infers. So include your city in the question, and run the test signed out, or with memory off. For each answer, open Sources, when available, and log these four things:
- The exact question and the city you included
- Every business named, in order
- Every source cited, with its URL
- Whether your business or any page of yours appears
Sort the cited sources by type
The sources cited next to your competitors show at least part of what ChatGPT drew on for that answer. Put each one in a bucket: the competitor's own site, a review platform, a directory or listing, a third-party comparison or best-of page, a news or industry publication, or a forum. Then count the buckets across all your questions.
Read the count as evidence, not as a ranking rule. Because citations can be incomplete, a source missing from the list may still have shaped the answer. Even so, a pattern across ten questions beats a guess. If competitor pages dominate, test your own pages first. If third-party lists and review pages dominate, the gap to test first is probably off your site.
Work on the bucket where you are absent, then test again
Pick the bucket your test points to and change one thing there at a time, so the next run can show whether it made a difference. OpenAI's published documentation does not name any of these as ranking factors, so treat each one as a change to test, keep every change factual, and start with the largest gap rather than doing a little of everything.
- Their own site is cited and yours is not: open the page that answers the question and compare it with yours. Does yours state the answer in its first two sentences, name the service, the city and the price range if you publish one, and cover the follow-up questions a buyer asks? Our earlier post on why a business is invisible to ChatGPT covers structure.
- Third-party lists and comparison pages name them and not you: find out whether each page accepts submissions, corrections or editor contact. Give the editor a factual reason you belong. Do not offer paid placements or link swaps.
- Review platforms feature them: check how many recent reviews you have where your customers actually look, and ask happy customers directly. Our guide on turning happy customers into 5-star reviews covers the asking, and Reputation & Trust covers the wider work.
- Directories list them: make your name, address, phone number and description identical on your site, your Google Business Profile and every listing. OpenAI's published documentation does not say whether this affects ChatGPT, but it removes conflicting facts about you.
Measure again, and expect uneven results
Run the same questions again after a month and compare the logs. Apart from the roughly 24 hours it gives for robots.txt changes, OpenAI's published documentation does not say how quickly a change on your site or elsewhere shows up in answers, and it says placement is not guaranteed, so treat any promised timeline with suspicion. Expect to appear for some questions and still miss others.
You can also watch referral traffic. OpenAI's publisher FAQ says publishers who allow OAI-SearchBot can track visits from ChatGPT in analytics tools such as Google Analytics, because ChatGPT adds utm_source=chatgpt.com to referral URLs. A mention only becomes a visit when someone clicks, so expect fewer visits than mentions, but the trend is real data. Our GEO service page describes a larger version of this loop, 20 buyer questions run three times each on ChatGPT, Perplexity, Google AI Mode and Copilot, with a citation map of who is cited instead of you, if you would rather have it run for you.
What to do this week
- Check robots.txt and your CDN or host rules for OAI-SearchBot, and confirm OpenAI's published IP ranges are not blocked.
- Write ten questions a buyer would ask, each with your city in it.
- Run each question three times, signed out or with memory off. Log the businesses named and the sources cited.
- Sort every cited source into the buckets above and count them.
- Pick the one bucket with the biggest gap and change one page or one profile there.
- In Google Analytics, filter for sessions from chatgpt.com so you have a baseline before anything changes.
You cannot control what ChatGPT chooses to say. You can control whether its crawler can reach you and whether the facts about you are consistent, and you can work on whether the places it cites know you exist. Do those three things and measure the result.