Brand Visibility in AI Search: 5 Signs You're Invisible

Brand visibility in AI search is whether ChatGPT, Perplexity, Gemini, Google AI Overviews and Claude name your brand, describe it correctly and link to you when someone asks a question you should answer. You are invisible when they skip you, get your facts wrong, or cannot reach your site at all, and each of those has a different cause and a different fix.

Below are the five signs, the check behind each one, and the three mistakes that cause most of them. Every fix here is something you control. None of them is a trick.

The five signs at a glance

First, know which AI answer you are looking at

An AI answer comes from one of two places. Either the model answers from what it learned in training (its memory), or it searches the web first and answers from the pages it fetched. OpenAI says ChatGPT may search the web automatically when your question would benefit from current information, so the same question can go either way.

Memory has a cutoff date. As of October 2026, OpenAI's chat-latest model, which OpenAI says points to the latest Instant model used in ChatGPT, lists an April 20, 2026 knowledge cutoff. Anything you launched or changed after that is not in the model's memory.

This split decides what you can fix. Nothing you publish this quarter changes a memory answer, while access and published facts change search answers as soon as the crawlers read your pages again.

If an answer shows no source links, assume it came from memory, but you cannot be certain.

Sign 1: AI names your competitors, never you

Ask ChatGPT and Claude the question a buyer would ask: "What are the best [your category] tools for [use case]?" Use a logged-out browser window (Claude requires an account, so use one with no chat history about your brand and note that in your log), because the same OpenAI article says that with memory on, ChatGPT may use your saved memories when it rewrites a search query. Then ask ChatGPT again with Search switched on: type / in the message box and pick Search.

• Missing both times: neither memory nor search finds you. Work through Signs 2, 3 and 5 first.
• Missing from memory, named with search on: the model's training data predates you or barely mentions you. Common for newer brands, and the search route is working.
• Named from memory, missing with search on: the search route is broken. Go straight to Sign 2.

What to do

Be honest about this one: no technique makes a model remember you sooner. A critical survey of 45 GEO studies found that no reviewed technique shows a stable, longitudinal, cross-platform causal effect on organic discoverability. What you can do is keep the search route open and make sure everything published about you is correct.

One trap to avoid: allowing GPTBot does not get you cited. GPTBot is OpenAI's training crawler, and OpenAI states that each of its crawler settings is independent of the others (sources under the crawler table below). Blocking it does not remove you from ChatGPT search either.

Sign 2: Perplexity and ChatGPT search never cite your site

AI search engines read the web with their own crawlers, and each one has its own gate. If the crawler cannot fetch your pages, the engine cannot cite them, however good they are. Here is what each major bot does, in the vendors' own words:

Sources: OpenAI's crawler documentation, Perplexity's crawler documentation, Anthropic's crawler help article and Meta's web crawler documentation.

The check

Open yoursite.com/robots.txt. Look for any group that names one of the search bots above with Disallow: /, and for a catch-all User-agent: * group that disallows everything. Either one shuts the search crawlers out.

robots.txt is only half the gate. OpenAI's ChatGPT search help article says that to be eligible, you must allow OAI-SearchBot and confirm that your host or CDN allows traffic from OpenAI's published searchbot IP addresses. A firewall rule that returns 403 to the bot removes you as completely as a Disallow line, and robots.txt still says allow.

To test it, request a page with the bot's user agent from a terminal (Terminal on Mac; on Windows, open PowerShell and type curl.exe instead of curl, because plain curl is a different command there):

curl -sIL -A "Mozilla/5.0 (compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexitybot)" https://www.yoursite.com/pricing

If you see a 301 or 308 first, the last status code is the one that counts.

A 200 is a pass. A 403, 429 or a challenge page is a lead worth chasing, not proof: some firewalls block fake bots on purpose, and your request comes from your own IP.

Confirm in your CDN's logs, filtered to the bot's published IP ranges.

If your site is on Cloudflare, check its AI crawler settings now. Since September 15, 2026, Cloudflare's "Block" and "Block on pages with ads" settings apply to mixed-use crawlers, including Googlebot, Bingbot and Applebot, so a setting meant to stop training can now cost you search too. Our guide on whether you should block AI crawlers on Cloudflare walks through each option.

Sign 3: Google AI Overviews cite your competitors, not you

Google's rule is short. To be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet, and Google says there are no additional requirements. So the failures are specific and checkable:

1. Not indexed. Run the page through URL Inspection in Search Console.
2. Snippet blocked. A noindex, nosnippet or max-snippet:0 rule, in the page's robots meta tag or in an X-Robots-Tag HTTP header. Google states that nosnippet and max-snippet also limit what can be used as input for AI Overviews and AI Mode.
3. Facts hidden by data-nosnippet. Often added around a cookie banner and copied, by template, onto the content that matters.
4. Opted out in Search Console. The Search generative AI control is set to include by default, but a child property inherits its parent's setting, so one change at the domain level can remove a whole subfolder.

Ranking well helps less than most people assume. Ahrefs found that 38% of URLs cited in AI Overviews also appeared in the top 10 (March 2026).

Also keep in mind that Google's same page says AI Overviews often don't trigger. If no Overview appears for a query, that is Google's choice, not your failure.

Sign 4: AI gets your facts wrong

This is worse than being invisible, because a wrong answer reaches the buyer as fact. The usual errors: an old price, a free plan you closed, a feature you dropped, a former CEO, or your brand blended with a similarly named company.

Start with a fact sheet: founding year, headquarters, headcount, CEO, ownership, current prices and plans, main features. Ask each AI platform about each one and mark every answer right, wrong or outdated. Then trace each wrong answer to its source:

• It cites your own page. Your site states the fact wrong somewhere, or states it two ways. Find every place the fact appears (pricing table, FAQ, footer, old blog posts, schema) and make them agree.
• It cites someone else's page. A directory, review site or old article is out of date. Ask them to correct it.
• It cites nothing. It is probably old training data. You cannot edit a model's memory. Publish the correct fact in plain HTML so search answers get it right, and check again after the next model release.

Our step-by-step guide on fixing ChatGPT's wrong information about your company covers the correction process in detail.

Sign 5: Your facts appear only after JavaScript runs

Open your pricing page, press Ctrl+U (Cmd+Option+U on Mac) to view the source, and search for your price. If it is not in that source, most AI crawlers never see it.

Vercel's crawler study found that ChatGPT's and Claude's crawlers fetch JavaScript files but don't execute them. Googlebot and Applebot do render JavaScript, which is why a page can look fine in Google and be blank to everyone else.

The fix is a developer task: render prices, product names and key claims on the server, so they are in the HTML before any script runs. Check the pages that carry facts first: pricing, product, About and comparison pages.

The three mistakes behind most invisible brands

Mistake 1: Assuming your Google rankings cover AI search

Ranking on Google tells you nothing about whether OAI-SearchBot, PerplexityBot or Claude-SearchBot can read your site. Each AI platform runs its own crawler behind its own gate, and even inside Google most AI Overview citations come from outside the top 10. Check each platform separately.

Mistake 2: Blocking AI wholesale, or letting your CDN do it for you

A blanket "block AI" rule usually catches the search crawlers along with the training crawlers. If what you want is "don't train on my content, but do show it in answers," say exactly that, bot by bot:

# Keep AI search and answer engines
User-agent: OAI-SearchBot
User-agent: PerplexityBot
User-agent: Claude-SearchBot
User-agent: Meta-WebIndexer
Allow: /

# Opt out of model training (Meta-ExternalAgent also covers Meta product indexing)
User-agent: GPTBot
User-agent: ClaudeBot
User-agent: Meta-ExternalAgent
Disallow: /

# Blocks Gemini training AND Gemini Apps grounding: only if you accept leaving Gemini answers
User-agent: Google-Extended
Disallow: /

If your file already has a User-agent: * group with Disallow lines, copy those lines into this group too: a bot follows only its most specific group and ignores the rest. If nothing blocks these bots now, you do not need this group at all.

Blocking Google-Extended does not affect Google Search: Google states it does not impact a site's inclusion in Google Search. It does cut Gemini Apps grounding, so decide on purpose.

Mistake 3: Checking once, logged in, on one platform

One check in your own ChatGPT account tells you what ChatGPT says to you, on that day, in whichever mode it chose. Answers change with each model release and each crawl.

Keep a fixed list of prompts, run it logged out on every platform you care about, and repeat it on a schedule. Our AI visibility monitoring routine takes 15 minutes a week.

What will not make an invisible brand visible

Much of what is sold as "AI optimization" adds markup or rewrites pages. The evidence for it is weak or negative, so do it only after the five checks above, if at all.

A 30-minute check you can run today

1. Write your fact sheet: eight to ten facts a buyer would ask about.
2. Ask ChatGPT, Claude and Perplexity three category questions and three brand questions, logged out where the platform allows it. Run ChatGPT once with Search on.
3. Open your robots.txt and confirm the search bots in the table above are allowed.
4. Request your pricing page as PerplexityBot and OAI-SearchBot. Note any 403 or challenge page.
5. View the source of your pricing page and find the price in it.
6. Check your top pages for noindex, nosnippet, max-snippet:0 and data-nosnippet.
7. In Search Console, confirm the pages are indexed and the generative AI control is set to include.

If you want a quick baseline first, the free AI visibility checker shows what ChatGPT and Google AI Overviews say about your brand. For the full list of fixes in priority order, use our GEO checklist.

Frequently asked questions

Why is my brand not showing up in ChatGPT?

Either ChatGPT answered from memory and its training data predates or barely covers your brand, or it searched and could not use your site. Ask the same question with Search switched on. If you appear only with search, your access works and the gap is the model's memory; if you never appear, check that OAI-SearchBot is allowed in robots.txt and that your CDN lets OpenAI's published IP addresses through.

How do I check my brand's visibility in AI search?

Ask the questions your buyers ask on each AI platform, in a logged-out window, and record whether you are named, whether the facts are right and which sites are cited. Then check the technical side: robots.txt per bot, a request to your site as each bot, and your pages' snippet controls. The first tells you what AI says; the second tells you why.

Does schema markup or llms.txt improve AI visibility?

No credible evidence says so. Google states structured data isn't required for generative AI search and that Google Search ignores AI text files such as llms.txt, and a matched-control test found schema slightly lowered AI Overview citations. Schema still has a job in Google rich results; it is not an AI visibility lever.

How long after a fix will my brand show up in AI answers?

Access and fact fixes reach search-based answers once the crawler fetches your pages again, which you can see in your server logs. Answers from a model's memory change only when a new model with a later cutoff ships. Nobody can promise a date for either, and no one controls whether a given answer uses search.

To check your crawler access, snippet controls and answers against your facts across the AI platforms (all six on paid plans; the free plan covers two), run a 99Visibility audit.