What Does ChatGPT Say About My Brand? How to Check It
To see what ChatGPT says about your brand, ask it the questions your buyers ask in a logged-out window, once as it answers by default and once with Search switched on, and score every answer against a written sheet of your own facts. The two runs can disagree, because ChatGPT answers from what it learned in training unless it decides to search the web.
That is the honest answer to "what does ChatGPT say about my brand": it depends on the mode, the day and who is asking. Below is the full check, step by step, what each kind of answer tells you, and what a manual check cannot see.
Why ChatGPT can give two different answers about you
OpenAI says ChatGPT may search the web automatically when your question would benefit from current information. When it does not search, it answers from its training data. Those are two different sources with two different fixes:
| Answer from memory | Answer with search | |
|---|---|---|
| Where it comes from | Training data, frozen at the model's knowledge cutoff | Pages fetched at answer time, through OpenAI's crawler and search partners |
| How to tell | No source links | Source links and citations in the answer |
| What changes it | A new model release | Your pages and third-party pages, once they are crawled again |
| What you control | Nothing this quarter | Crawler access and the facts you publish |
The cutoff matters more than most brands realize. 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. A price change, rebrand or new product after that date does not exist in the model's memory.
Search does not see your words either. OpenAI says ChatGPT search typically rewrites your query into one or more targeted queries that it sends to its search partners. The prompt you type is not the query that finds, or misses, your site.
Step 1: Write your fact sheet (10 minutes)
You cannot spot a wrong answer without a written right one. List the facts a buyer could ask about, and the date each was last true:
- One sentence on what you do and who it is for
- Founding year and headquarters
- CEO, and the year they started
- Owner, parent company and any acquisitions
- Headcount range
- Every current plan and price, with the date you last changed them
- Your five main features
- What you do not do: retired plans, dropped features, markets you left
The last line catches the errors that hurt most: a free plan you closed, a product you discontinued, a partnership that never happened.
Step 2: Set up a clean session
Open a private browser window and do not sign in (Claude requires an account, so for Claude use one with no history about your brand). OpenAI says web search works for people who are not signed in, and that with memory on, ChatGPT may use your saved memories when rewriting a search query. Your own account has learned about your brand from your own chats, so it is the worst place to test.
Location matters too. The same OpenAI article says ChatGPT may use your IP address to estimate your country, state or city. If you sell in several countries, run the check from each one or note where you ran it.
Step 3: Ask six kinds of questions
| Type | Example prompt | What it tests |
|---|---|---|
| Identity | "What is [Brand]? Who owns it and who runs it?" | Founding, CEO, ownership facts |
| Pricing | "How much does [Brand] cost? Is there a free plan?" | Prices and plans, the most time-sensitive facts |
| Category | "What are the best [category] tools for [use case]?" | Whether you are named at all when nobody asks for you |
| Comparison | "[Brand] vs [Competitor] for [use case]: which is better?" | Positioning, and whether you are confused with a rival |
| Problem | "How do I [the job your product does]?" | Whether you come up before the buyer knows your category |
| Reputation | "What do customers say about [Brand]?" | Tone, and which review sites it leans on |
Two of each type gives you twelve prompts. Keep the wording fixed so you can compare runs over time.
Step 4: Run each prompt with search off and on, three times
- Default run. Paste the prompt into a new chat. If the answer shows no source links, treat it as the memory answer.
- Search run. Start a new chat, type / in the message box and choose Search, then paste the same prompt. OpenAI documents this, and also a "Search the web" option on the regenerate control.
- Repeat. Run each version three times in fresh chats. The same prompt can come back different, so count how often you are named rather than trusting one answer.
There is no documented switch that guarantees ChatGPT will not search. That is why the absence of source links is your best signal for a memory answer, and why it is only a signal.
Step 5: Score every answer
Log each run in a sheet. One row per answer:
| Date | Prompt | Mode | Named? | Facts right? | Wrong fact | Sources cited | Competitors named |
|---|---|---|---|---|---|---|---|
| 2026-10-16 | Pricing #1 | Search | Yes | No | Says free plan exists | g2.com, yoursite.com/blog/2024-launch | 2 |
| 2026-10-16 | Category #1 | Memory | No | n/a | n/a | None | 5 |
From that log you get two numbers that matter: mention rate (answers that name you, divided by all answers) and accuracy (facts stated correctly, divided by facts stated). Keep them separate for memory and search runs. Our guide to GEO metrics explains how to report them without overstating what they mean.
What each kind of answer tells you
| What you see | Likely cause | What to do |
|---|---|---|
| Not named from memory, named with search | The model's training data predates you or barely covers you | Nothing changes memory this quarter. Keep the search route open |
| Not named even with search | OpenAI's crawler is blocked, or no page of yours answers the question | Check robots.txt and your CDN (below) |
| Wrong fact, cites your own page | Your site states it wrong, or two ways | Make every instance on your site agree |
| Wrong fact, cites another site | A directory, review site or old article is out of date | Ask the site owner to correct it |
| Wrong fact, no sources | Old training data | Publish the correct fact in plain HTML; recheck after the next model release |
| Your brand blended with another | A similar name, and thin or inconsistent pages about you | State your full name, domain and category together on your About page and profiles |
| Awards, clients or features you never had | Fabrication, or a third-party page that says so | Log the exact wording; trace and correct any page that states it |
How to fix what ChatGPT gets wrong
Open the search route
OpenAI's crawler documentation says sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers, and recommends allowing both the bot in robots.txt and requests from its published IP ranges. Check both. A robots.txt that allows the bot does nothing if your firewall returns 403 to it.
User-agent: OAI-SearchBot
Allow: /If your robots.txt already has a User-agent: * group with Disallow lines, copy those lines into this group too, because a bot follows only its most specific group. If nothing blocks OAI-SearchBot now, you do not need to add anything.
GPTBot is a separate decision. It is OpenAI's training crawler, and OpenAI states each of its crawler settings is independent of the others: blocking GPTBot does not remove you from ChatGPT search, and allowing it does not get you cited.
State every fact once, correctly, in plain HTML
Find every place your site states each fact on your sheet: pricing page, FAQ, footer, old blog posts, schema markup. Make them agree, then view the page source and confirm the fact is there.
Vercel found that ChatGPT's crawler fetches JavaScript files but does not execute them, so a price that loads by script is invisible to it.
Correct the third-party pages it cites
If the wrong answer cites a directory, comparison article or review profile, that page is the problem. Update the profiles you control and ask other site owners for corrections. Our guide on fixing ChatGPT's wrong information about your company covers the outreach.
Do not chase the memory answer
No technique makes a model relearn your brand before its next release. Spend the effort on what the search route can read, and recheck memory answers when OpenAI ships a model with a later cutoff.
What a manual check cannot tell you
- Which mode a real buyer got. ChatGPT decides per question whether to search.
- The query it actually ran. It rewrites prompts before searching, and OpenAI does not show you the rewrite.
- What signed-in users see. Memory and location change the search query.
- Your "rank" in the answer. OpenAI's whole disclosure is that it ranks results "using multiple factors" and that placement is not guaranteed. Position in a generated answer shifts run to run and has nothing to validate it against, so do not track it.
- Why you are missing. The answer shows the symptom. Only a check of robots.txt, your CDN and your page source shows the cause.
- The other four platforms. Perplexity, Gemini, Google AI Overviews and Claude each read the web differently.
The free AI visibility checker shows what ChatGPT and Google AI Overviews say about your brand, with a score, in one run. A full 99Visibility audit queries ChatGPT and Claude through their APIs with web search on, labels each answer as searched or from memory so you can tell the two apart, and checks the answers against your facts.
Does ChatGPT send you traffic? How to see it
OpenAI's FAQ for its ChatGPT Atlas browser stated that ChatGPT automatically includes utm_source=chatgpt.com in referral URLs (archived October 4, 2026). Atlas has since been retired, the FAQ now redirects, and no current OpenAI page repeats it, so check your own GA4 source list for chatgpt.com rather than relying on the tag. Google Analytics 4 also groups visits from assistants such as ChatGPT in an AI Assistant channel, added on May 13, 2026.
Read that number as a floor. A memory answer has no links, so it sends no clicks however often it names you. Traffic measures a fraction of what ChatGPT says about you, which is why the answer log matters more than the analytics.
For the platform-by-platform picture, see our ChatGPT optimization lesson, and if ChatGPT does not name you at all, work through the five signs of poor brand visibility in AI search.
Frequently asked questions
How do I check what ChatGPT says about my company?
Write a fact sheet, then ask ChatGPT identity, pricing, category, comparison, problem and reputation questions in a logged-out window. Run each prompt by default and again with Search turned on, three times each, and log whether you are named, whether the facts are right and which sources it cites.
Why does ChatGPT say it doesn't know my brand?
It is answering from memory, and its training data either predates your brand or barely mentions it. Ask again with Search turned on. If it finds you then, your site is reachable; if it still does not, check that OAI-SearchBot is allowed in robots.txt and that your CDN does not block OpenAI's published IP addresses.
Can I change what ChatGPT says about my brand?
You can change what its search answers read: allow OpenAI's search crawler, publish each fact correctly in plain HTML, and correct third-party pages it cites. You cannot change what the model already learned; that only changes when OpenAI releases a model with a later cutoff.
Does blocking GPTBot remove my brand from ChatGPT?
No. GPTBot is OpenAI's training crawler, and OpenAI states each crawler setting is independent. ChatGPT search inclusion depends on OAI-SearchBot, so you can block GPTBot from training and still appear in ChatGPT search answers.
How often should I check what ChatGPT says?
Weekly for the handful of prompts that matter most, such as pricing and your main category question, and a full run monthly. Recheck immediately after you change prices or after OpenAI announces a new model. Our AI visibility monitoring routine keeps that to 15 minutes.
To see what ChatGPT, Google AI Overviews, Google AI Mode, Gemini, Perplexity and Claude say about you, checked against your facts and with each platform's mode stated (all six on paid plans; the free plan covers ChatGPT and Google AI Overviews), start a 99Visibility audit.
Want more of our research in Google Search and AI Mode? Add us as a preferred source.
See what AI says about your brand
Your free audit asks ChatGPT and Google AI Overviews about your brand. Paid plans add Gemini, Perplexity and Claude, and show every answer that contradicts your facts.
Start for Free- ChatGPTMentioned, 2nd of 5 brands
- Google AI OverviewsNot mentioned
- PerplexityMentioned and cited as source [1]
- ClaudeSays $29/mo. Your pricing page says $49/mo.
Related Articles
ChatGPT Wrong Information About My Company? How to Fix It
There is no form that edits what ChatGPT says about you. Here is what works: find where the wrong fact came from, fix the source, and use each platform's report route.
Brand Visibility in AI Search: 5 Signs You're Invisible
Five signs AI search is skipping or misdescribing your brand, the check behind each one, and the three mistakes that cause most of them.
How to Rank in ChatGPT: What Decides Whether It Mentions You
There is no rank inside a ChatGPT answer. Here is what decides whether ChatGPT mentions your brand, with search on and off, and the four checks to run today.