How to get mentioned by ChatGPT and other AI assistants

October 10, 2026·8 min read

To get mentioned by ChatGPT, your brand has to be described on the pages ChatGPT reads when people ask about your category, and described the same way each time. Those pages are mostly roundups, review sites, comparison articles and forum threads written by other people, so most of the work happens off your site. Behind that, make the brand easy to identify as one entity, with one name, Organization markup, sameAs links and a Wikidata item, and give every source the same specific line about what you do and who it's for.

A mention and a citation are different wins. ChatGPT can name you in a "best X for Y" answer without linking any page of yours. This post covers how to earn the mention. Getting your own pages used as linked sources is a separate job, covered in how to get cited by ChatGPT.

Brand mentions in AI vs citations

A mention is the answer naming your brand. A citation is a link to a page, shown as the source for part of the answer.

	Mention	Citation
What it is	Your brand named in the text of the answer	A link to a page, shown as a source
Example	"Acme Rota suits clinics with under 20 chairs"	A source link to a review site's comparison page
Where it comes from	Training data, plus pages read during a search	Pages read during a search
Who writes the source	Mostly other sites	You, when the cited page is yours
How you notice it	Asking the questions yourself	Asking, plus visits tagged `utm_source=chatgpt.com`

The two often come apart. Ask for the best scheduling tool for a dental clinic and ChatGPT may search, read a review site's comparison page and a Reddit thread, then write a shortlist. The review site and Reddit get the citations. The brands on the shortlist get the mentions, and the shortlist is what the reader acts on.

How to get mentioned by ChatGPT with and without web search

With web search, ChatGPT can mention you as soon as the pages it retrieves describe you. Without it, the model can only name brands it learned about before its training cutoff. OpenAI says ChatGPT chooses whether to search based on what you ask, so the same brand question can go either way.

Search is the faster route, and the one you can work on this month.

Training is slower. Anthropic publishes the cutoff for each Claude model and warns that a model may not know about events after it. A brand launched after the cutoff does not exist for that model, and a brand described on a handful of pages before it may not have stuck. To be named from memory, a brand has to be described widely and consistently, on many independent sites, before the next model's cutoff. Treat AI brand mentions that come from training as a lagging result of the work below.

Your own site is one voice in that mix. Blocking GPTBot keeps your pages out of OpenAI's training data, per OpenAI's crawler docs, and does nothing about what other sites say about you.

How to get recommended by ChatGPT for "best X for Y" questions

Find the pages AI answers already cite for your category's buying questions, then get onto those pages with accurate information.

List 10 to 20 questions a buyer would type. Include the category with a qualifier ("best rota software for dental clinics"), head-to-head comparisons and "alternatives to [your biggest competitor]".
Ask each one in ChatGPT with search on, several times, and write down every source link and every brand named. AI Citation Finder does the counting at scale. Enter a topic and it lists the 25 pages cited in the most ChatGPT and Google AI Overviews answers, plus the sites cited most.
Group the sources into roundups and comparison articles, review platforms, forum threads and competitors' own pages.
Work through them, starting with the sources that appear in the most answers.

Roundups and comparison articles

Email the author with what they need to add you. Say who the product is for, what it costs and how it differs from the tools already listed, and offer an account to try it. Five lines is plenty.

If the article lists you with an old price or a dropped feature, ask for the correction first. It is an easier yes than a new listing, because it fixes the author's own error, and it stops answers that cite the page from repeating the wrong fact.

Review sites

Claim your profile on each review platform in your sources and make its description, category and pricing match your site. Ask customers for reviews once they have had a result, and ask all of them the same way. Google's Organization markup docs list a review-site profile as a valid sameAs link, so add the profile to your markup too.

Reddit and forum threads

Threads where someone asks "what should I use for X" get read by search engines and by AI. OpenAI's partnership with Reddit, announced in May 2024, gives it access to Reddit's Data API to bring Reddit content into ChatGPT.

Answer existing threads rather than starting new ones. Use a personal account, say that you work for the company, answer the question that was asked, and name a competitor when it fits the asker better. Many subreddits limit self-promotion or ban it, so read the rules first. One useful answer in a thread that already ranks does more than a dozen posts nobody upvotes.

Make your brand one entity: name, markup, sameAs and Wikidata

Entity consistency means every page that describes you uses the same name and points to the same site, so the mentions add up to one brand instead of several near-matches.

Start with the name. Pick the exact form, "Acme Rota" and not "AcmeRota" or "Acme Rota Software", and use it in page titles, og:site_name, Organization markup, social profiles, review listings and press boilerplate. Put the registered company name in legalName and old names or spellings in alternateName, so they point at the same entity instead of competing with it.

Then mark up the home page or About page, where Google's Organization docs recommend putting it. The Q number below is a placeholder for your Wikidata item.

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Acme Rota",
  "alternateName": "AcmeRota",
  "legalName": "Acme Rota Ltd",
  "url": "https://acmerota.example",
  "logo": "https://acmerota.example/logo.png",
  "description": "Acme Rota is shift scheduling software for dental clinics with 2 to 20 chairs.",
  "sameAs": [
    "https://www.wikidata.org/wiki/Q00000000",
    "https://www.linkedin.com/company/acmerota",
    "https://www.g2.com/products/acme-rota"
  ]
}

Google wants the logo at least 112 by 112 pixels and accepts several sameAs URLs. Each one should open a profile that exists and uses the same name.

Wikidata comes last because not every company qualifies. Its notability policy accepts an item for a clearly identifiable entity that serious, publicly available references describe, so you need independent coverage first. Then create an item with your official website (property P856) and the sources that describe you, and add its URL to sameAs.

Give AI one quotable description of your brand

Write one or two sentences that say what you are, who it is for and one concrete fact that sets you apart, then use them everywhere you describe yourself. A "best X for Y" prompt always has a Y. When your description names the same Y, you are an obvious candidate.

Vague:    Acme Rota is a modern scheduling platform for growing teams.
Specific: Acme Rota is shift scheduling software for dental clinics with
          2 to 20 chairs. It books hygienists around chair availability
          and costs $49 per location per month.

The specific version names a category, an audience and a price, so an answer can repeat any of them. Use it word for word on:

your home page and About page
the description in your Organization markup
LinkedIn, Crunchbase and every review profile
directory listings and press boilerplate
the notes you send to roundup authors

When the price changes, change it everywhere in the same week. An old price on a review profile keeps getting repeated long after your own site is fixed.

What not to do

Do not fake reviews, astroturf Reddit or write your own Wikipedia article. Each goes against a published rule or guideline, and each can leave a public record that AI reads next to your name.

Fake reviews. The FTC's rule on fake reviews, in force since October 21, 2024, bans fake reviews, paying for reviews that must express a particular sentiment, and insider reviews that hide the connection. The FTC can seek civil penalties against knowing violators.

Reddit astroturfing. Reddit's site rules ban content manipulation, including spam and vote manipulation. When users spot sockpuppet accounts praising your product, the thread exposing them becomes one more page about your brand for search and AI to read.

Your own Wikipedia article. Wikipedia's conflict of interest guideline strongly discourages editing articles about your own company and requires paid editors to disclose who pays them. Suggest changes on the article's talk page and let independent editors decide.

How to measure LLM brand visibility

Ask the same buying questions on a schedule and count how often each AI engine names you, in what position and next to which competitors. The same question gets different answers on different runs, so ask each one several times and track the rate.

AI Answer Visibility runs a customer's question through ChatGPT, Perplexity and Gemini with web search on, and reports whether each answer names or cites you and which brands it names instead. For the process, read AI brand monitoring on checking what AI says about you, and share of voice in AI answers on comparing your mention share with competitors.

Check your brand entity signals

Brand Entity Checker reads your home page and tests the signals that tie mentions of your brand to your site. It checks that one name appears across your Organization and WebSite markup, og:site_name, the application-name tag, the web app manifest, the llms.txt title and the home page title, and it accepts the names you list in alternateName. It confirms that Organization markup has a url and a logo that loads and meets Google's 112-pixel minimum, opens up to ten sameAs links to check each profile exists, and looks up whether a Wikidata item lists your site as its official website.

Each problem comes back as a finding with the fix, and a run costs 8 credits. The checker does not create a Wikidata item or track mentions on other sites. To see what ChatGPT, Perplexity and Gemini say about your brand, use Brand Narrative Check.