Why does ChatGPT say it has no information about my company?

ChatGPT has two ways to know about you. The first is parametric knowledge: what the model learned in training, frozen at a cutoff date. The second is live search: pages fetched at answer time and placed in the model's context.

"No information" means both came up empty for that question. Either your company isn't in the training data, or search didn't run, or search ran and couldn't reach or read your pages.

That gives you a checking order. Start with the cheapest question and work down:

1. Does your company postdate the model's knowledge cutoff?
2. Did search actually run for this answer?
3. Can OpenAI's search crawler get past your robots.txt and your firewall?
4. Are your key facts in the HTML before JavaScript runs?
5. Is your name shared with something better known?

Treat each as a check. None of them is a finding until you've looked. If you're starting from zero, our guide on finding out what AI says about your company covers how to run a clean baseline.

The knowledge cutoff

OpenAI's model pages list a February 16, 2026 knowledge cutoff for the GPT-5.6 family, and August 31, 2025 for the Instant path ChatGPT uses (as of October 2026). Routing between them changes what "ChatGPT knows" by months. If you launched, rebranded or got press after those dates, the trained model can't know it. Nothing you publish today reaches a retrieval-off answer until a future model is trained. Our page on whether ChatGPT can know a new company goes deeper on this.

Whether search triggered

ChatGPT search is model-discretionary. The model decides per question whether to search. When it doesn't, you get the cutoff answer. The tell is citations: an answer with no source links was almost certainly not grounded in a live search.

What to do: ask again and request sources, or phrase the question so it plainly needs current information. If a searched answer names you and an unsearched one doesn't, your problem is the cutoff, not your site. Test logged out with memory off, because OpenAI says saved memories can shape how ChatGPT rewrites a search query.

robots.txt, per bot

OpenAI runs separate crawlers, and its bots page says each setting is independent. OAI-SearchBot controls ChatGPT search: sites opted out of it "will not be shown in ChatGPT search answers." GPTBot controls training use only. A common error is a blanket AI block that removes search eligibility by accident. Allowing GPTBot does not help search.

A minimal rule that allows search looks like this:

User-agent: OAI-SearchBot
Allow: /

The firewall layer

OpenAI's help page says eligibility needs both: allow OAI-SearchBot in robots.txt and confirm your host or CDN allows traffic from OpenAI's published search IP addresses. A bot-management rule can return a 403 or a challenge page while robots.txt says yes. Since September 15, 2026, Cloudflare blocks Training and Agent crawlers by default on ad-displaying pages for new domains and free-tier customers, while leaving Search allowed. Check your own settings rather than assuming. The AI Crawler Checker requests a page as each AI crawler and shows which ones your firewall turns away.

JavaScript rendering

Vercel's crawler study found ChatGPT's crawlers fetch JavaScript files but don't execute them, and can't read client-side rendered content. If your company description, pricing or product names load only after scripts run, compare the raw HTML with the rendered page. Anything missing from the raw HTML may not exist for OpenAI's crawlers.

OpenAI says ChatGPT search rewrites your question into one or more targeted queries before sending them to its search providers. If your company shares a name with a bigger brand, a place or a common word, those queries can return the other entity. The model then reports what it found, which may be nothing about you.

Say you run a hypothetical bakery called Summit in Denver. A bare question about "Summit" competes with every Summit online. Add your category and city to the test question and see if the answer changes.

A few edge cases to keep in mind:

• Third-party mentions. When engines cite other sites, being named on those sites is associated with appearing in answers. That's an association, not a promised effect.
• A wrong answer is a different problem. If ChatGPT describes you incorrectly rather than saying nothing, see hallucination vs outdated answers.
• Things that won't help. Schema markup, llms.txt files and "AI-friendly" rewrites have no credible evidence of getting a company named.

Work the list top to bottom and stop when you find a block:

1. Note your launch and rebrand dates against the model cutoffs.
2. Re-ask with sources requested, logged out, and compare grounded and ungrounded answers.
3. Read your robots.txt for OAI-SearchBot specifically.
4. Test whether your CDN or firewall lets OpenAI's search crawler through.
5. Diff raw HTML against the rendered page for your key facts.
6. Test with your category and location added to the question.

Then put one plain sentence on a fetchable page that states what you are, what you sell and where. A specific stated fact gives the engine a sentence it can quote. It's the only on-page signal with positive experimental support, and it still guarantees nothing.

Be clear about control. You control access, rendering and what your pages say. You don't control when search fires, how OpenAI ranks results, or when the next model is trained. For the difference between fixing trained knowledge and fixing search, read fixing ChatGPT errors: trained knowledge vs search.