Why your brand does not show up in AI answers: five causes

A marketing manager asks this question after a buyer says that ChatGPT did not name the brand. This guide names five causes, with a test and a fix for each. Our own project is the example, with every number from its last 28 days. For the moves that follow a diagnosis, read how to get mentioned in ChatGPT answers. For what the numbers can and cannot tell you, read how accurate AI visibility numbers are.

Why does my brand not show up in AI answers?

A brand does not show up in AI answers when one of five conditions holds. The AI crawlers cannot read its site. The lists that the answers cite leave it out. No page of its own answers the question. The tool does not count a spelling of its name. Or the question names no company at all. The first three causes are about the brand. The last two are about the measurement.

In the 28 days to 30 October 2026, on 7 AI platforms, our own project stored 8,953 answers to 47 prompts, all in English. 14 of the answers named our brand. 13 of the 14 came from one prompt, which named us in 13 of its 191 answers. 45 of the 47 prompts did not name us once.

Do AI crawlers read your site?

An AI crawler is the program a platform sends to read a page before it answers or cites it. A robots.txt file can refuse one by name. FullReach AI reads the robots.txt and the homepage of your site and of each monitored rival every day. It gives each of 31 crawlers a verdict: allowed, blocked, or blocked with exceptions.

On 30 October 2026, the audit of our own site and of the five rival sites found no blocked crawler on any of the six. Each site allows all 31 crawlers with one Allow rule. Our homepage returned 12,576 characters of text to a plain read. So a block is not our cause. Our guide on checking whether AI crawlers can read your website holds the crawler list and the manual checks. It also holds the change to make when robots.txt blocks a crawler.

Do the cited lists include you?

The answers that name your rivals and not you cite pages, and those pages show where you are missing. A losing answer is an answer that names a monitored competitor and not your brand. For each losing answer, FullReach AI stores the pages it cites. The All Sources page sorts those pages by domain. The Actions page turns each cited best-of page into one item of work: get listed there.

In the 28 days to 30 October 2026, 4,395 of our answers were losing answers. The cited best-of list with the most losing answers appeared in 482 of them. Our guide on getting mentioned in ChatGPT answers holds the moves that get a brand onto such a list.

Home›Actions

Example

ActionAction kindPlatformsLosing answers

Get a profile on review sites21 / 2 done

Create a profile on g2.comReview site2 of 738

Create a profile on capterra.comReview site2 of 727Done

Publish on sites and forums10 / 1 done

Take part in reddit.com/r/crm, openlyForum1 of 721

Answer what competitors’ sites are cited for10 / 1 done

Publish a page that answers what hubspot.com is cited forCompetitor’s site1 of 714

Does a page of yours answer the question?

An answer cites a page of yours when that page answers the question the buyer asked. The Own Pages page lists every page of your site that the answers cite or read. In the 28 days to 30 October 2026, 24 of our 8,953 answers cited a page of ours. 67 answers cited or read one. 10 answers cited a page of ours and named other brands only. Among the cited domains, fullreach.ai ranked 174th, with 45 citations.

The 45 citations went to 9 pages. The homepage took 10. Two comparison pages took 13. Four guides took 18. The pricing page and the facts page took 4. The 10 answers that cited us without naming us show the gap. The platform read the page and used it, and it still named other brands. The Actions page names the page to publish or to improve, with the prompt behind it. The guide on getting mentioned holds how to write that page.

Does the tool count your name?

A brand counts in an answer only when the answer writes its name, or a spelling that the project holds. A platform may write a name with a space, without one, or as the domain. An answer that writes a spelling the project does not hold counts as an answer without you. The test: open the Competitors page, open your brand's row, and read "Seen in answers". It lists the spellings the answers used and the project does not hold, each with its count. Add the ones that are your brand. In the 28 days to 30 October 2026, the answers wrote our name as Fulreach in 5 answers, a spelling the project does not hold.

The same page lists the brands that the answers name and nobody monitors. In our window, Semrush appeared in 2,471 of 8,953 answers, and the project does not monitor it. The fix on the web is one name. Write it the same way on your site, your profiles and your listings, so that a platform meets one spelling.

Does the question name anybody at all?

A prompt whose answers name no company measures nothing about visibility. FullReach AI marks such a prompt "names nobody" when no answer named a company of any kind on its last 7 collection days. In the 28 days to 30 October 2026, none of our 47 prompts carried the mark. But 5 prompts named a monitored brand in 8 or fewer of their 192 answers, and all five ask how or why.

The prompt "What is the best AI visibility tool?" named a monitored brand in 183 of its 192 answers. The fix is a different prompt, one that asks for tools, providers or alternatives. Our guide on building a prompt set from Search Console queries turns buying queries into such prompts. Our guide on which prompts to track measures how much more often a request for tools names a brand.

In which order do you test?

Test the two measurement causes first, then the three brand causes.

1. Read the "names nobody" marks on the All Prompts page. Reword or drop a silent prompt.
2. Read "Seen in answers" under your brand on the Competitors page. Add your spellings.
3. Read the Crawlability page. Change robots.txt when it blocks a search or user crawler.
4. Read the cited pages of your losing answers. Ask for a place on the lists.
5. Read the Own Pages page. Publish the page that answers the prompt.

Then read the same prompts again after a full window. Our guide on how accurate AI visibility numbers are states what a 7-day and a 28-day window each hide. A single check by hand reads one answer. Our one winning prompt named us in 13 of its 191 answers, so most single checks miss us.

What to do next

1. Run the two measurement tests on your prompts first.
2. Fix the one brand cause your tests found, and write down the day of the change.
3. Read the same prompts after a full 28-day window, and compare the two windows.

For how FullReach AI collects the answers, read the methodology.