Usually because ChatGPT cannot find your pages when it searches, while your competitors' pages, or the directories and comparison articles that list them, are right in front of it. In the answers we measure every week, the causes show up in a fairly steady order: you are not in the index the assistant searches, your crawlers are blocked, you have no page that answers the buyer's question, and your competitors sit on the sites the assistant cites.
This is the field now called generative engine optimization: improving how often a brand or page appears in answers that AI systems retrieve, summarise and present, which works differently from a list of ten blue links. Below is how to diagnose each cause, in the order we see them most often.
Is ChatGPT actually searching when it names your competitors?
Start here, because it changes what you fix. In a test on 23 September 2026, we put 20 German business questions to a model. It answered 10 of them without searching at all and cited nobody. Where it did search, it cited primary sources such as laws, ministries, the OECD and large accounting firms on 9 of 10. Firms were named only on "who should I hire" and "how do I choose" questions.
So look at the question your buyer types. If it is a "which provider" or "best tool for" question, the assistant tends to search and name businesses, and the causes below apply. If it is a general how-to question, there may be no company named at all, yours or anyone's.
Also check whether the gap is stable. In the same test, two runs of the same question named two different sets of firms. A critical survey of GEO research reports the same pattern from commercial audits: citations vary across repeated runs of the same query, and different engines cite different ecosystems of sources. One screenshot of a competitor being named proves little. We count a business as named for a question if any of three asks names it, and we compare weeks only on the same questions and the same number of asks.
Is your site in the index the assistant searches?
This is the most common cause we find, and the one owners least expect, because their site is "on Google".
On 19 October 2026 we asked assistants for the exact titles of 18 articles, each older than three weeks, published on three client sites. ChatGPT found 1 of 18. Claude found 7 of 18. Google had indexed 29 of 30 articles of one of those sites. Being in Google's index did not mean an assistant could find the page.
Size of the crawlable surface matters more than polish. One site we looked at had complete markup: server rendering, Article, FAQPage and Organization JSON-LD, and a sitemap. With only 75 URLs, it was still absent from answers on its own article's exact topic. A site with about 190,000 crawlable URLs was crawled constantly.
You can see the difference in live retrievals too. Agents such as ChatGPT-User, OAI-SearchBot and Perplexity-User fetch a page while an assistant is answering a real person. In one fortnight the edtech platform we work with received 1,918 such fetches on content pages. A client blog of a few dozen articles received 56, spread across 24 different articles.
What to check:
- Bing Webmaster Tools as well as Google Search Console. Look for pages marked "discovered but not crawled".
- Your sitemap: does it list every page you want found, and is it submitted to both?
- IndexNow, so new and changed pages are announced instead of waiting for a recrawl.
- Ask ChatGPT and Perplexity for the exact title of one of your pages. If they cannot find it, they cannot cite it.
We walk through these tests in how to check if ChatGPT can see your website.
Are you blocking AI crawlers without knowing it?
The second cause is a door you closed yourself. We know because we did it.
On itgrows.ai, a robots.txt rule meant to hide a single redirect page, /b, was written as a prefix. Robots.txt rules match prefixes, so from 11 August to 21 October 2026 that one line also closed /blog, /business and every article in our business blog to all crawlers. Bing Webmaster Tools had listed our business page as "discovered but not crawled" since 11 August 2026. Google had none of the business pages indexed. We changed the rule to /b$ on 21 October 2026.
These crawlers are not a rounding error. Across the full retained nginx logs of pickaclass.com from 28 July to 11 August 2026, ClaudeBot made 19,272,125 requests, Meta-ExternalAgent 3,854,076, GPTBot 2,744,985, Applebot 2,655,555 and Amazonbot 1,170,268, against 52,292 from Googlebot. If your robots.txt, CDN bot settings or firewall turn them away, your competitors are the only ones left to read.
Read your robots.txt line by line, test a few deep URLs against each rule, and check your server logs for the user agents above. Our guide on whether your website is blocking AI crawlers covers the common traps.
Do you have a page that answers the buyer's question?
Once the assistant can reach your site, it still needs something to quote. Most small business sites describe the company: services, team, contact. Buyers ask questions: which option fits my situation, what does it cost, how do I compare providers. If no page on your site answers that question directly, the assistant uses a page that does, and that page often belongs to a competitor.
Google documents that its AI Overviews and AI Mode can use query fan-out, issuing several related searches across subtopics to assemble one response. Each sub-question is a chance to be found or skipped. A single "Services" page competes for none of them.
The content itself matters. The GEO paper by Aggarwal et al. found that adding citations, quotations from relevant sources and statistics to web content can raise a source's visibility in generative-engine answers by over 40%. Pages that state facts an assistant can lift and attribute give it something to use.
A local example shows the split. On 20 October 2026, for an orthopaedic practice in Bonn, the four assistants named the practice in 12 of 80 answers to 20 patient questions: Claude 5 of 20, Gemini 4, ChatGPT 2, Perplexity 1. "Who should I choose" questions named practices. Questions about costs, risks and treatments named none. Those answers still draw on sources, so a clear page on those topics is your chance to be among them.
The practical move: list the questions your buyers ask before they choose, and publish one clear page per question on your own domain. More on this in how to get ChatGPT, Claude, Gemini and Perplexity to recommend you.
Why do competitors appear on the sites the assistant cites?
The fourth cause is about the rest of the web. When an assistant answers "who should I use", it often builds the list from directories, comparison articles and vendor blogs. If your competitors are on those pages and you are not, they get named.
In the Bonn measurement, the directory jameda.de was cited in 12 answers, and competing practices' own websites were the most cited sources of all.
Our own category shows the same thing. On 21 October 2026 we put 20 English and 20 German questions about getting a business into AI answers to ChatGPT, Claude, Gemini and Perplexity, three times each. ItGrows was named in 0 of 160 answers and none of our pages was cited. English answers most often cited firstpagesage.com and llmpulse.ai (8 answers each), help.openai.com (7), and blog.hubspot.com, ahrefs.com and searchengineland.com (6 each). The tools named most in English answers were Profound (12 answers), Semrush (9), Otterly.AI and Peec AI (8 each) and Ahrefs (7). Most cited sources were the blogs of GEO agencies and tool vendors themselves.
This is not always a verdict on quality. The GEO-Bench paper treats manipulation of LLM product rankings as a recognised concern for fairness and information integrity, which tells you these rankings respond to signals, not just merit. That cuts both ways: the signals are ones you can work on.
Collect the cited URLs for your questions. Get listed on the directories that keep appearing, and publish your own comparisons and buyer guides so there is a page of yours in the same pool.
How do the four causes compare?
| Cause | How to check | What fixes it |
|---|---|---|
| Not in the index the assistant searches | Ask ChatGPT and Perplexity for an exact page title; check Bing Webmaster Tools and Search Console | Complete sitemap, submission to Bing and Google, IndexNow, more crawlable pages |
| Crawlers blocked | Read robots.txt for prefix rules; look for GPTBot, ClaudeBot, OAI-SearchBot in server logs | Narrow rules (for example /b$ instead of /b), allow AI user agents in CDN and firewall |
| No page answers the question | List buyer questions and look for a page that answers each one directly | One page per question, with facts, sources and figures that can be quoted |
| Competitors on cited directories and comparisons | Record which URLs the assistant cites for your questions | Directory listings, your own comparison and buyer-guide pages |
What should you do next?
First, find out where you stand on your buyers' questions, not on your brand name. The free check at https://www.itgrows.ai/ generates five buyer questions from your homepage and gets answers from ChatGPT and Perplexity in 25 to 40 seconds.
Then work down the causes in order: index, crawler access, answering pages, cited sources. Measure on the same questions, three asks per assistant, every week. It is cheap to do yourself: with API prices of October 2026, a report of 20 questions across 4 assistants with 3 asks costs us about $5.40.
Keep expectations honest. In about two weeks of logs from three of our course websites, human visitors arriving from AI assistants numbered 1 from Copilot and 0 from ChatGPT, Perplexity, Gemini, Claude, Meta AI and others. Assistants mostly answer in place, so being named matters more than clicks. Some clients report outcomes anyway: FreightSizer, a B2B SaaS for freight dimensioning, reported 2 new B2B customers in one month, both from ChatGPT. And on 20 October 2026 Perplexity cited an article we had published on a client's site, a German company-formation advisory.
If you would rather not run this yourself, ItGrows publishes question-answering articles on your domain, fixes crawler access and indexing, and reports weekly how often ChatGPT, Claude, Gemini and Perplexity name you, for $499 a month (plus a one-time $499 setup, free with 3 months or a year). Nobody can guarantee that an assistant will name a business, and we do not.
Frequently Asked Questions
Why does ChatGPT name different companies each time I ask?
Answers vary between runs: in our tests, the same question asked twice named two different sets of firms. Research on GEO reports the same variation across repeated queries. Judge visibility over several asks and several weeks, not one screenshot.
Is ranking on Google enough to be mentioned by ChatGPT?
No. In our 19 October 2026 test, Google had indexed 29 of 30 articles of one client site, yet ChatGPT found only 1 of 18 article titles across three sites. Check Bing Webmaster Tools and test exact titles in the assistants themselves.
Will schema markup make ChatGPT mention my company?
Markup alone did not do it in what we have seen. A site with full JSON-LD, server rendering and a sitemap but only 75 URLs was absent from answers on its own article's topic. Crawler access, indexing and pages that answer real questions come first.
How quickly will ChatGPT mention me after I fix robots.txt?
There is no fixed timeline, because crawlers have to revisit and the pages have to be indexed again. We fixed our own robots.txt rule on 21 October 2026 and track the effect with the same weekly questions. Submitting your sitemap and using IndexNow helps announce the change instead of waiting.