How to get your brand mentioned by ChatGPT, step by step

How to get your brand mentioned by ChatGPT

You get your brand mentioned by ChatGPT by being the brand that many trusted pages already name for your buyer’s question. ChatGPT mostly repeats what the web agrees on. Your job is to be part of that agreement.

That sounds vague, so this guide breaks it into eight steps. Each one points to the public evidence behind it, and says where that evidence is weak. None of it involves paying ChatGPT, because there is no way to do that for the answer itself.

OpenAI does sell ads now, but it says "Ads do not influence the answers ChatGPT gives you". Ads are labelled as sponsored and shown apart from the answer (OpenAI, Testing ads in ChatGPT). A mention inside the answer has to be earned.

A mention and a citation are not the same thing

A mention is your brand’s name in the text of the answer. A citation is a link to a page that the answer used. You can have either one without the other.

For most businesses the mention matters more. A buyer who asks for "the best accounting software for a small charity" reads the names in the answer. Fewer of them open the links.

If you want your own pages linked as a source, that is a related but different job, covered in our guide to getting cited by ChatGPT.

What the evidence says drives mentions

The best public evidence says mentions across the web and on YouTube track AI visibility more closely than links or site size. It is correlation, not proof of cause, and Ahrefs says so itself.

Ahrefs took 75,000 brands and checked millions of AI answers for their names, using its own Brand Radar tool. It then compared how often each brand was named with ordinary search measures (Ahrefs, 75k brands studied). The Spearman correlation it reports runs from 0 (no link) to 1 (move perfectly together).

Two findings stand out for a smaller brand. First, Ahrefs found that ChatGPT showed weaker links to classic authority measures than Google’s AI products did. It suggested ChatGPT "may be the best entry point" for a brand with modest search volume and backlinks. Second, the number of pages on a site barely mattered. Publishing more pages is not the same as being named more.

Ahrefs also warns that "correlation isn't causation". Brands that are famous get mentioned everywhere and also get named by ChatGPT. That does not prove one more mention will move you. It does tell you where the brands that get named are present.

Getting your brand mentioned, step by step

Start by finding out where you stand, then work on the places ChatGPT reads, then measure again. Here are the eight steps in order.

1. Write down 20 to 50 questions your buyers ask before they choose a product like yours. Most should not contain your brand name.
2. Ask each question in ChatGPT several times, with search on, in fresh chats. Record every brand named and every page cited.
3. For each question, list the pages ChatGPT cited. These are the pages that decide who gets named today.
4. Check whether those pages mention you. If a roundup names five rivals and not you, that page is your first target.
5. Get onto those pages honestly: send the author facts, a trial account or a correction. Ask, never pay for a hidden placement.
6. Earn mentions where Ahrefs saw the strongest link: videos, reviews, community answers and trade press that name you for the problem you solve.
7. Make your own site say, in plain sentences, what you sell, who it is for, where you operate and what it costs.
8. Ask the same questions again in four to eight weeks, in the same way, and compare the counts.

Step 3 matters most. Seer Interactive found that when questions used words like "best" or "top", ChatGPT most often cited affiliate, news and aggregator sites (Seer Interactive). For a buying question, the shortlist in the answer is often a summary of those lists. Our explainer on why AI assistants recommend your competitor goes through the other common reasons.

For a buying question, ChatGPT’s shortlist is often built from other people’s lists. Get onto the lists it reads, and the answer tends to follow.

Step 7 is the one you fully control. A page that says "We make payroll software for restaurants in Ontario, from $40 a month" gives a model something exact to repeat. A page that says "We empower teams to unlock growth" gives it nothing. Both are made-up examples.

What to say on your own site

Your own pages should state the facts a model needs to place you: what you are, who you are for, and how you differ. Write them as plain sentences a model can lift whole.

There is some research support for writing this way. The GEO study, published at KDD 2024, found that adding statistics, quotations and cited sources to a page raised how much of it appeared in AI answers, while stuffing in keywords did not help (Aggarwal and others). It tested a research setup and Perplexity, not ChatGPT, so take it as a direction.

Keep the facts the same everywhere. If your site says one price, a directory says another and an old press release says a third, a model has three answers to choose from. Fix the stale copies you can reach.

What not to do

Avoid the shortcuts that either do not work or put the brand at risk. Each item below is either contradicted by the evidence or works against the very trust that gets a brand named.

• Do not mass-produce pages. Ahrefs found almost no link between the number of pages on a site and how often the brand was named.
• Do not stuff your pages with the question’s words. In the GEO study, keyword stuffing gave little or no gain, and on Perplexity it did about 10% worse than leaving the page alone.
• Do not post fake reviews or pose as a customer in forums. A model that reads a thread can also read the replies calling it out, and you lose the trust you were trying to borrow.
• Do not pay for a placement that hides the payment. If a roundup ranks you because you paid, say so, or the list stops being something a reader or a model should trust.
• Do not judge the work on one screenshot. One answer is a sample of one, and the next run can differ.

The honest route is slower, and it is the only one that lasts. The brands that get named are the ones other people keep writing about for the same problem. Being worth writing about is still the root of it.

It also helps to know which questions are worth the effort. A question that names you is easy to win and tells you little. A question about your category, where the buyer has not yet heard of you, is where a mention changes who gets the sale. Spend your time on those.

Why one check is never enough

ChatGPT writes each answer fresh, so you can be named once and missed the next time. Count mentions over many questions and many runs before you decide anything.

Location changes answers too. OpenAI's help page says ChatGPT may use a general location from the IP address, and saved memories, when it rewrites a question for search (OpenAI help). Two people asking the same thing can see different shortlists.

Citations also fade. Profound, which sells AI visibility software, measured 883,000 cited pages for a year and found that "the median page loses half its citation share within 11 days of peaking" (Profound). That study is about citations, not mentions, but the lesson carries over: a good week is not a lasting position.

• Ask each question at least three times before you count it as a win or a loss.
• Keep questions that name your brand apart. Being named when the buyer typed your name is not the same achievement.
• Compare the same questions month to month. A new question set gives a new baseline, not a trend.

How to track mentions without doing it by hand

You can track this in a spreadsheet, and for 20 questions that is fine. Past that, a tool saves the hours, as long as it shows you the answers behind its numbers.

Agentcites is one such tool, and we make it. It reads your website, writes the questions a buyer would ask, and puts them to ChatGPT, Perplexity, Gemini and DeepSeek every week. It keeps each answer word for word and records which brands were named. It counts questions that name your brand apart from those that do not, and it learns your real rivals from who keeps winning. The features are listed on the features page, and how we measure explains every count.

Other tools do similar work, some with more assistants or a free tier. Our comparison pages set them side by side with the sources for each fact, and say plainly that we wrote them.