Why does ChatGPT not mention my company at all?

Why does ChatGPT not mention my company at all?

ChatGPT leaves a company out when its search crawler cannot read the company's pages, when those pages never say plainly what the company does and for whom, or when the sources it cites for the category do not mention the company. Company facts that differ between the website and its profiles make the problem worse, and you can check every one of these causes yourself in an afternoon.

How do I find out which cause it is?

Check in this order. A crawler that cannot get in makes every later check pointless.

  1. Open yoursite.com/robots.txt in a browser. Look for a group that names OAI-SearchBot, PerplexityBot or Claude-SearchBot, and read the group that starts with User-agent: * as well. A line that says Disallow: / in a group that applies to a crawler blocks that crawler from the whole site. OpenAI says a site that blocks OAI-SearchBot will not be shown in ChatGPT search answers, and that a robots.txt change takes about 24 hours to reach its systems.
  2. Open the event log of your firewall or CDN (in Cloudflare, the Security events view) and search for requests from OAI-SearchBot, PerplexityBot and Claude-SearchBot. A firewall can refuse a crawler that robots.txt invites. OpenAI and Perplexity both publish the IP addresses their crawlers use, so your firewall can allow them.
  3. In Google Search Console, open Settings and read the Search generative AI setting for your domain property and for any child property. A property set to Exclude gets no traffic or impressions from AI Overviews and AI Mode, and a child property can inherit Exclude from its parent without anyone touching it.
  4. Give your homepage and About page to someone who does not know the business, and ask them to write one sentence on what you do, for whom and where. If they cannot, an assistant has little to repeat either.
  5. List your company name, legal name, description, address, phone number and founder. Write down how each appears on your website, in the Organization markup (paste a page into the Schema Markup Validator to see it), on LinkedIn, on your Google Business Profile and in the directories of your field. Where the details differ, an assistant has to guess which one is right.
  6. Write five questions a buyer asks before choosing a company like yours, and one question that names your company. Ask them in ChatGPT, Perplexity and Claude on three different days, and save each answer with its date and the pages it cites. Answers change between runs, so look for a pattern across days and assistants before you conclude anything.

Steps 1 and 5 have a shortcut. The free Visibility Mesh scan checks eight AI crawlers one by one against your robots.txt and scores whether your brand and products are stated the same way in the data AI systems read. Keep the free scan result as your dated starting point.

What does each cause look like, and what fixes it?

The AI search crawlers cannot fetch the pages

Each assistant runs a search crawler that is separate from its training crawler. OpenAI documents OAI-SearchBot for ChatGPT search and GPTBot for training, Perplexity documents PerplexityBot, and Anthropic says that disabling Claude-SearchBot may reduce a site's visibility in Claude's search results. A rule written to stop AI training often stops the search crawlers too. You may block training and still allow search. What you want to rule out is blocking both by accident. Write down a decision for each crawler and name the person who owns it, the way a security team does for every account on a system.

The site does not say plainly what the company does and for whom

An assistant can only repeat what a page states. A homepage that says "we help brands grow" gives it nothing to repeat, while "we repair commercial refrigeration for restaurants in Phoenix" names a service, a customer and a place in one sentence. Our AI Legibility Framework calls this AI Readability: whether the key facts are stated plainly or buried in marketing prose.

Company facts conflict across the site and its profiles

A business that appears under two spellings, two descriptions or two addresses gives an assistant two candidates for one company. The fix is one spelling for the legal name, the brand name and each product, used on the site, in the markup and on every profile you control. Our guide to entity recognition covers the markup side.

The sources assistants cite for the category leave the company out

Assistants that search the web cite comparison articles, directories, review sites and trade press. When those pages list your competitors and not you, the answer follows them. Google says its AI features are rooted in its core Search ranking systems, so the pages that rank for your category's questions tend to shape the answer as well. Earning a place on those pages is the slowest part of the work, and the citation footprint guide shows where to start.

What does this look like on a real kind of website?

Take a bookkeeping firm on Squarespace that serves independent restaurants. Its robots.txt allows every crawler, so step 1 passes. Step 4 fails at once: the homepage headline says "Numbers that work for you", and the stranger guesses the firm is a bank. The About page calls the firm by a shorter name than the one on LinkedIn and on its Google Business Profile, so step 5 fails too. The fix is two sentences at the top of the homepage that say the firm does bookkeeping and payroll for independent restaurants in its city, and one spelling of the name everywhere. After that, step 6 is repeated on the same dates each month to see whether the answers start to name the firm.

Which fixes can I make myself, and which need help?

You can allow OAI-SearchBot, PerplexityBot and Claude-SearchBot in robots.txt if you want to appear in AI search, and decide separately about training crawlers such as GPTBot. You can set the Search generative AI setting to Include, rewrite the first paragraph of your homepage and About page, and put the same description on the profiles you control. Asking real customers for reviews on the platforms your market uses is also yours to do. A developer changes firewall and CDN rules, edits robots.txt where the platform generates it, and corrects Organization markup written by the theme. What Visibility Mesh can change for you on each platform is set out on the platform support page.

How do I tell whether a fix worked?

A robots.txt change is confirmed by fetching the file again and running a rescan. A firewall change is confirmed in the event log, where crawler requests should now get through, and markup is confirmed in the Schema Markup Validator. Presence in answers is confirmed only by asking the same questions again on fixed dates. Even then, assistants change on their own, so a new answer suggests the work helped without proving it.

Questions people ask

How do small businesses show up in ChatGPT at all?

A small business shows up in ChatGPT when the sources ChatGPT reads describe it clearly, and a small business can be described as clearly as a large one. Allow the AI search crawlers in robots.txt, make the homepage and About page say what you do, for whom and where, keep the same name and description everywhere, and ask real customers for reviews.

Why does ChatGPT know my product exists but never recommend it?

Describing a product and recommending it are different steps for ChatGPT. Asked about your company by name, it can describe what it read about you. Asked for the best option in a category, it picks the products its sources tie most closely to that use, so publish the use cases and the type of buyer plainly and get listed where your category is compared.

How soon does ChatGPT notice after I unblock its crawler?

OpenAI says a robots.txt change takes about 24 hours to reach its systems. When an answer starts to use your pages after that depends on the question and on the other sources ChatGPT chooses, so record the answers to your fixed questions before the change and compare them on fixed dates afterwards.

Sources

We read every source below on September 26, 2026.