Why would AI say my company closed when it's still open?
99Visibility Research · Updated October 21, 2026 · 5 min read
Short answer
AI says a company closed when its sources say so or when it has no current sources at all. The usual causes are stale training data, third-party pages that list you as closed, confusion with a similarly named business that did close, or your own site being unreachable to AI crawlers. Check which sources the answer cites first, then your crawler access, then fix the pages you control.
Key takeaways
- If the answer cites no sources, it probably came from the model's frozen training data, which nobody can edit this quarter.
- If the answer cites sources, one of those pages is usually the origin of the closure claim, and that page is the thing to fix.
- A business with a similar name that really did close is a common source of confusion, so make your name, location and category unambiguous on your own site.
- If AI crawlers cannot fetch your site, engines have only other people's pages to go on.
- Nobody can force a model to change its answer, but you can remove the blocks and correct the sources it draws on.
The short answer, and why it happens
An AI answer that says you closed is repeating something. It comes from one of two places: the model's training data, or pages it retrieved for this question. The fix depends on which.
There are four likely causes:
- Stale training data. With no live search, a model answers from what it learned before its knowledge cutoff. OpenAI's model docs list a Feb 16, 2026 cutoff for the GPT-5.6 family and Aug 31, 2025 for the Instant path ChatGPT uses. An old closure of one location, a past rebrand, or a temporary shutdown can live on in that memory.
- Third-party pages. A directory, review site, old news story or map listing marks you as closed. When retrieval is on, the engine may quote it.
- Name confusion. Another company with a similar name did close, and the model merges the two.
- Your site is unreachable. If AI crawlers can't fetch your pages, the engine can't see that you're open. Only outside sources remain.
A fifth possibility is plain fabrication. Telling a hallucination from an outdated answer matters, because the fixes differ.
Diagnose in this order
Work from the cheapest check to the most involved one.
- Record the exact answer. Note the platform, the date, the wording, and whether it shows citations. Ask the same question a few times. Retrieval fires per query and at the model's discretion, so answers vary.
- Open every cited source. If the answer links pages, find the sentence that says closed. That page is your origin. Common culprits are a closed second location, an outdated directory listing, or a news story about a different business.
- If there are no citations, suspect training data. Ask whether the closure claim predates the model's cutoff. Did you move, pause, rename or shut a branch before then?
- Search for name collisions. Look up your exact name plus "closed" or "shut down". The same pattern explains why AI still recommends a competitor that shut down.
- Check crawler access. Confirm robots.txt allows the search bots (OAI-SearchBot, PerplexityBot, Claude-SearchBot), your firewall doesn't return 403 to them, and your status is in the HTML before JavaScript runs.
- Check index status per engine. Look at Google Search Console and Bing Webmaster Tools separately.
For a fuller workflow, the hallucination triage playbook walks through each branch.
What the evidence says about access
Access problems are the cause most in your control, and the evidence on them is strong.
- robots.txt rules are per bot. OpenAI's crawler docs say sites opted out of OAI-SearchBot "will not be shown in ChatGPT search answers". Blocking GPTBot affects training only. OpenAI states each setting is independent.
- A firewall can override robots.txt. OpenAI's help centre says eligibility requires allowing OAI-SearchBot and confirming your host or CDN allows its published IP addresses. A 403 from bot protection is a silent failure.
- JavaScript can hide your content. Vercel found ChatGPT and Claude crawlers fetch JavaScript files but don't execute them. If your hours or "we're open" banner loads client-side, those crawlers never see it.
- Indexing gates Google's AI features. Google says a page must be indexed and eligible for a snippet to appear as a supporting link in AI Overviews or AI Mode. In Otterly's 2026 experiment, Google's AI surfaces almost stopped citing two sites once Google deindexed them, while Copilot kept citing them from Bing.
Also check for stray noindex, nosnippet or data-nosnippet on your homepage or contact page. These keep content out of AI answers.
The AI Crawler Checker on 99Visibility (request your page as each AI crawler) shows which bots your firewall turns away.
Fixes you control, and what you don't
In your control
- State your status plainly on a fetchable page. Put a sentence like "Acme Plumbing is open and serving Austin" with current address and hours in server-rendered HTML. A specific, quotable fact gives the engine a sentence it can use.
- Correct the source pages. Update or claim the directory and map listings that say closed. If a closed branch caused it, mark that branch as closed and the main business as open. When engines cite other sites, being described correctly there is the route into the answer.
- Disambiguate your name. Use your full name, city and what you do consistently across your site and listings. See also wrong location or contact details, which shares the same fixes.
- Remove access blocks found in the diagnosis.
Not in your control
- What a model learned before its cutoff. That changes only when a new model ships.
- Whether an engine searches for a given question, and which pages it picks.
No fix guarantees a corrected answer. Re-check after changes, platform by platform.
What to do next
Start by seeing the problem clearly. Ask the main engines the questions a customer would ask, note which ones say closed, and record their sources. Our guide on finding out what AI says about your company covers how to run those checks consistently.
The free checker on this site gives a quick snapshot of how AI answers up to three buyer questions about you. Then fix the highest-leverage item first: a blocked crawler or a cited wrong source usually beats rewriting copy.
Frequently asked questions
Can I ask ChatGPT or Google to correct the closure claim directly?
Feedback buttons exist, but nothing published shows they change answers for everyone. The dependable route is fixing what the engines read: your own pages, the third-party listings they cite, and crawler access. If the claim comes from training data, it persists until a newer model ships, though answers that use live search can reflect your corrections sooner.
Will adding schema markup or an llms.txt file stop AI saying I'm closed?
There's no evidence it will. Google says structured data isn't required for its generative AI features and that Google Search ignores llms.txt. Ahrefs' matched test found AI Overview citations fell 4.6% after pages added schema. Existing markup does no harm, but a plain, fetchable sentence stating you're open is the more useful fix.
Does blocking GPTBot make ChatGPT think my business closed?
No direct link is documented. GPTBot controls training use only, and OpenAI says blocking it is independent of search. Blocking OAI-SearchBot is different: OpenAI says those sites won't be shown in ChatGPT search answers. Then ChatGPT relies on other pages or its training data, which may be out of date.
Why does one AI say I'm closed while another says I'm open?
Each engine uses different models, cutoffs and indexes. Google's AI features draw on Google's index, Copilot on Bing's, and ChatGPT and Perplexity use their own crawlers and search providers. One may retrieve a stale directory while another finds your site. Check each platform separately rather than trusting a single blended result.
What if AI says it has no information about my company at all?
That is a related but different problem: the engine found nothing usable rather than something wrong. The access checks are the same. Our page on [why ChatGPT says it has no information about your company](/questions/chatgpt-no-information-about-company) covers that case.
Sources
This page is AI-assisted. It was written from a verified evidence base and checked against our editorial policy. Engine behaviour changes often; facts are as of October 21, 2026.