Am I visible in ChatGPT? You can find out in an hour, for free. This AI visibility check has twelve questions in four areas: access, clarity, evidence and presence. Score each from 0 to 2, add them up, and read your band out of 24.
Most owners find out the hard way. A prospect says they asked ChatGPT, and a competitor’s name came back. A brand can sit on page one and still be absent from the written answer above the results.
This visibility check needs no paid tools. You need your website, your robots.txt file, a notebook and a few AI assistants. It will not replace an expert read, but it shows where the gaps are and whether you need one.
Businesses ask ChatGPT about their own category once and draw a conclusion from one answer. A structured AI visibility check does more than one question can.
This guide is about brand visibility, meaning whether ChatGPT names your business to other people. Privacy of your own chats is covered near the end.
ChatGPT and other AI search engines name brands they can reach, understand and cross-check
The audience is large. OpenAI’s own research paper says that by July 2025 more than 700 million people used ChatGPT weekly, sending more than 2.5 billion messages per day (OpenAI and NBER, 2025). Some of those messages are your buyers asking who to trust.
When a ChatGPT user asks for a recommendation, ChatGPT can search the web and retrieve relevant pages. OpenAI describes answers with links to relevant web sources, shown under a Sources button (OpenAI, 2024).
The answer is built from pages the system can read and compare. AI answers are not a list of links, so traditional SEO alone does not decide who is named.
SEO aims to drive traffic and is measured by keyword rankings and organic sessions. Answer engine optimization aims to get you cited in AI answers, and is measured by brand mentions and share of voice.
So the check follows the chain from access to answer. Can AI crawlers reach your site, with open crawler access? Is it clear who you are? Do other sources say the same thing? And when someone asks, does the AI name you?
Our guide to what AI visibility is explains the causes in more depth.
Score each question 0, 1 or 2, and score “unknown” as 0
Score 2 only if the answer is true today and you could show it to a stranger. Score 1 if it is partly true, or true for some pages, platforms or people. Score 0 if it is untrue, untested or unknown.
“Unknown” scores 0 because a gap you have not seen is still a gap. Write each answer down with the date and repeat the check in three months. The difference between the two sheets is your real result.
Access: can AI crawlers reach and read your site?
ChatGPT is powered by a large language model, and a crawler can only cite what it can fetch. Accessible, relevant pages come first. If a system cannot fetch your pages, nothing else matters. These three questions reveal the cheapest faults.
1. Does your robots.txt allow the crawlers behind ChatGPT, Perplexity and Gemini?
Open yourdomain.com/robots.txt. Score 2 if search crawlers such as OAI-SearchBot and PerplexityBot are allowed and you chose that on purpose. Score 1 if you have never read the file. Score 0 if a blanket rule blocks them.
OpenAI says sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though they can still appear as navigational links. It also says a robots.txt change can take about 24 hours to apply (OpenAI, 2026).
GPTBot is a different crawler. OpenAI describes it as used to make its generative AI foundation models more useful and safe. Blocking GPTBot is a separate decision from blocking search, so read each rule by name.
A third agent, ChatGPT-User, acts when a person asks ChatGPT to visit a page. OpenAI says that because these actions are user initiated, robots.txt rules may not apply to it.
Bing matters too. Seer Interactive matched more than 500 ChatGPT citations across about 100 questions and found that 87 percent matched Bing’s top 20 organic results for the same question, against 56 percent for Google (Seer Interactive, 2025).
Seer treats this as a small sample, so use it as a prompt to check Bing Webmaster Tools, not as a rule.
Perplexity states that PerplexityBot is designed to surface and link websites in search results, and recommends allowing it if you want to appear (Perplexity, 2026).
2. Are your key pages indexed and eligible to show with a snippet in Google?
Google states that a page must be indexed and eligible to be shown in Google Search with a snippet to appear in its AI features, with no additional technical requirements (Google Search Central, 2026).
That covers Google’s AI Overviews and Google AI Mode. Check your service, pricing and contact pages in Google Search Console.
Score 2 if every page that matters is indexed and none carries a “nosnippet” or “noindex” tag by mistake. Score 1 if some are missing. Score 0 if you have not looked.
3. Can a reader get your key facts as plain text, without a click, a form or JavaScript?
Google advises making important content available in textual form. Open your service page with images and scripts switched off, or view the page source.
Score 2 if your services, location, prices or price guidance, and contact details appear as ordinary text. Score 1 if some sit inside tabs, PDFs or sliders. Score 0 if the facts exist only in images or behind a form.
Clarity: does every source describe you the same way?
AI models place a brand by comparing descriptions. If your own pages disagree, the system has no clear picture to repeat. The aim is to make your brand easy to identify.
4. Can you describe what you do, for whom and where in one sentence, and does your homepage say it?
Score 2 if the sentence exists, your homepage opens with it, and it names your category in the words buyers use. Score 1 if the idea is there but buried. Score 0 if your homepage leads with a slogan and the category is never stated.
5. Are your name, address, services and description identical on your website, LinkedIn page, Google Business Profile and review sites?
Open each. Score 2 if the facts match to the letter, including contact details. Score 1 if they match in substance but differ in wording, dates or spelling. Score 0 if they contradict each other, for example old services or a former address.
This builds entity clarity: the system recognizes one organization, not three. See entity in the Brand Library.
6. Does each important page open with a direct answer to a question a buyer would ask?
Take your three most important pages. Score 2 if each opens with a plain answer, followed by the detail, and has a short FAQ with common questions and updated dates. Score 1 if the answers are there but start after several paragraphs. Score 0 if the pages describe the company and never answer a question.
Structured data and schema markup can help with rich results. Google states there is no special schema.org structured data you need to add for AI features, but that markup should match the visible text on the page (Google Search Central, 2026). Fix the plain text first.
Clear structure and direct, factual wording make a page easier for a system to parse and cite than vague marketing language.
Evidence: do independent sources back up what you say?
Your own site is one voice. AI systems lean on claims that other sources repeat, in public web content and off-site discussions such as forums, news and review sites. This area asks who else is talking about you.
7. Can you name three independent sources that mention your brand?
Search your brand name in quotation marks, excluding your own domain. Score 2 if you find three or more trusted sources of brand mentions, such as trade media, industry bodies or news, that describe what you do. Score 1 if you find one or two. Score 0 if the only results are your own profiles.
8. Do your reviews and ratings say what you want to be known for?
Read your last twenty reviews. Score 2 if they name the services and strengths you want associated with you. Score 1 if they are positive but generic. Score 0 if there are few reviews or they raise problems you have not answered.
9. Do you publish original material that others can cite, such as data, a method, expert commentary or a clear definition?
Score 2 if at least one page of yours is something others link to or quote. Score 1 if your content is useful but could be found anywhere. Score 0 if you publish only news about yourself.
ChatGPT citations tend to go to pages that state something specific and checkable. Our piece on answer engine optimization covers how to write for that.
Presence: when businesses ask ChatGPT, do the AI answers name you?
Everything above predicts your brand’s visibility. This area measures it, one question at a time. Three things matter here: whether you are named, whether you are described correctly, and whether your relevant pages are cited.
10. When you ask ten buyer questions about your category, how often is your brand named?
Write ten questions your customers would ask without your name in them, such as “which [your service] providers in [your city] do you recommend?” Ask each in ChatGPT, Perplexity and Google Gemini, and check Google’s AI Overviews for the same search.
Score 2 if you appear in four or more of the ten across the platforms. Score 1 if you appear in one to three. Score 0 if you appear in none.
Answers are generated each time, so ask every question at least twice and count a mention only if it appears more than once. Log the date.
OpenAI says ChatGPT considers the full context of your chat (OpenAI, 2024). Start a new chat for each question so that earlier answers do not shape the next one.
11. When ChatGPT names you, is the description correct and are your pages cited?
Separate AI mentions from AI citations. Brand mentions are your name appearing in the answer. A citation is the system linking to your page as a source.
Score 2 if you are described correctly and your own pages are linked. Score 1 if you are named but described vaguely, or never linked. Score 0 if you are misdescribed, or not named at all.
12. Do you track your AI visibility, and does it change what you do?
Score 2 if each week or month you record your prompt results, check Google Search Console for clicks and impressions, and note branded search traffic and enquiries that mention AI tools. Score 1 if you look occasionally. Score 0 if you track nothing.
Google says sites shown in AI features are included in the overall search traffic in Search Console, in the Web search type of the Performance report (Google Search Central, 2026). You cannot split that traffic out, so keep your own prompt record too.
A spreadsheet with the same ten prompts is enough. Track your citation rate over time: the share of the ten prompts where a page of yours is linked. Share of voice compares your brand mentions with those of your competitors. Paid AI visibility tools automate the same job across platforms.
Free tools and alternatives support your AI visibility check but cannot replace it
Plenty of tools now offer a free AI visibility check. Most ask you to sign up, enter a website and return a score. Each uses its own prompts and scoring, so treat the score with care.
Use tools to confirm what the twelve questions found. If a tool says you are invisible and your own check says you are named in six of ten answers, trust the check and ask why they differ. One question run once is a guess. Ten questions run twice, logged by date, is evidence.
Your total points to one of three next moves
Add your twelve scores. Your total is your AI visibility score, out of 24. It is a method score, not a research benchmark, so measure yourself against your own past result and your closest direct competitors.
0 to 8: ChatGPT probably cannot place you
The likely faults are basic: blocked crawlers, unclear descriptions, or no independent mentions. The early fixes are cheap.
9 to 16: you are findable but not yet the answer
The pieces exist but do not line up. Look at where you scored 0, not at the total. Zeros in access or clarity are fixes on your own site. Zeros in evidence mean you need other voices, which is a distribution job: earned mentions and partnerships. See what content distribution is.
17 to 24: strong AI visibility, so protect it
You are named and cited. The risk is drift: a site rebuild that blocks crawlers, a changed address, or a competitor publishing better evidence. Repeat the check every quarter and after any redesign.
A self-check has limits that an outside read covers
You scored your own brand with your own bias. Ten prompts are a small sample, and answers move, so your presence is better read as a trend than a single result. A low score does not tell you which fault to fix first.
For a worked example, read how a tourism business gets found when travelers ask AI.
The Brand Diagnostic turns your scorecard into a verdict
The Brandesis Brand Diagnostic is one session with our strategists and a written read of where your brand stands: your position, the evidence behind it, and the one thing holding growth back, as a one-page verdict.
It costs P 7,500. If you book a Sprint, Discovery or Build within ninety days, the fee is credited in full, as set out on the plans page.
Bring your scorecard and your prompt log. If you scored under 17, book the Brand Diagnostic or start a project. Reaching the right people through owned, earned and paid channels is the work of distribution strategy.
Repeating the check every quarter is how your ChatGPT visibility improves
One score is a snapshot. Three scores are a trend. Keep your prompt log, re-run it each quarter, and fix the lowest-scoring area first.
For the thinking behind the questions, read our guide to GEO, along with the Brand Library entries for share of voice and generative engine optimization. Our free ebook, When AI Answers First, goes deeper on being the business a machine names.
Every Wednesday, Ground Truth sends one argument about brand decisions and one thing you can test in your own business that week. It is free, and you can leave in one click.
Being visible to people is a different question from keeping your own chats private
Some searches for this phrase are about privacy, so here is the short answer. Visibility in your own chats is under your control. A shared link lets anyone who has it view the shared content, so review a conversation before you send the link (OpenAI Help Center, 2026).
A Temporary Chat stays out of your chat history and is not used to improve OpenAI models while it remains temporary (OpenAI Help Center, 2026). Your settings decide whether other chats are saved or used for training.
ChatGPT is not a social network. Other people do not see your online status, a live presence indicator or your typing in a private chat. Brand visibility is the opposite problem: it is whether the public answers ChatGPT gives to other people name you.
FAQ
Am I visible in ChatGPT if my website ranks on Google?
Not necessarily. A high ranking helps, but ChatGPT builds answers from sources it can reach and cross-check. Run the ten buyer questions in Question 10 to see whether your brand is actually named.
How do I check my AI visibility for free?
Write ten buyer questions without your brand name, ask them in ChatGPT, Google Gemini and Perplexity, and check Google’s AI Overviews. Count how often you are named and cited, and log the date. That log is your AI visibility data.
What is a good AI visibility score?
There is no industry standard. On this check, 17 or more out of 24 is strong. Compare yourself with your own earlier result and with your closest competitors, using the same prompts.
Is AI visibility the same as SEO?
No. Traditional SEO measures where your pages rank in a list of links. AI visibility measures whether you appear inside the written answer. They overlap, because AI Overviews draw on indexed pages, but a high ranking does not guarantee a mention.
How often should I repeat the check?
Every quarter, and after any website rebuild, rebrand or change to your robots.txt.