Brand mentions and AI recommendations

A brand mention is any page, video, review, or post that names a company or product, with or without a link. For AI recommendations, the mentions that count most are on other sites: in one study, only 2.9% of 149,912 AI citations pointed to the brand's own website (opens in a new tab), and across 75,000 established brands, branded web mentions correlated at 0.66 to 0.71 with AI visibility (opens in a new tab) (vendor study). Mentions are a correlate, not a proven lever, so earn real ones and measure.

This guide defines brand mentions, shows where AI assistants find them and what the research does and does not show, and gives steps for earning and tracking mentions honestly. It belongs to AI search optimization, AEO HQ's guide to search engine optimization plus corroboration, facts, and measurement. Advice specific to ChatGPT is in how to get recommended by ChatGPT, and measurement methods are in how to measure AI visibility. Its sources were checked on September 27, 2026.

Scope and definitions

• Brand mention. Any page, video, review, forum post, or listing that names a brand. A linked mention includes a hyperlink to the brand's site; an unlinked mention does not.
• Brand, earned, and social sources. One study of AI citations sorted sources into three types: brand sites, "official brand sites that directly provide products or services"; earned sources, "independent media, review, or comparison sites"; and social sources, "social platforms, community forums, and user-generated content" (opens in a new tab).
• Mention, citation, and recommendation in an AI answer. A mention names the brand. A citation links a page as a source; the page is often not the brand's own. A recommendation names the brand as an option for the user's need. They are measured separately; see how to get recommended by ChatGPT.
• Branded anchor. In one vendor's definition, the visible, clickable text of a hyperlink that features the brand name (opens in a new tab).
• Brand monitoring. Tracking where a brand's name appears, on the web and in AI answers.
• Share of voice. A brand's mentions as a share of all mentions of a named set of competitors, across a fixed set of prompts and runs.
• Entity. A distinct thing, such as a company or product, that search systems and models try to keep apart from others with similar names.

Work to be named and cited in AI answers is called answer engine optimization (AEO). This guide covers the part of it that happens on other people's websites.

How mentions reach AI answers

Two routes: training data and search

An assistant can name a brand from what its model learned in training, its parametric knowledge, or from pages it finds by retrieval while it answers.

• Training. In controlled experiments, a model could reliably extract a fact only when its training data restated the fact in varied forms (opens in a new tab) (peer-reviewed). On an 18,000-question benchmark, GPT-4's accuracy fell from well-known to lesser-known entities and averaged 31% (opens in a new tab) (peer-reviewed). A brand named in many independent places is restated many times; a brand named only on its own site is not.
• Search. When assistants search, most of what they cite is about the brand, not by it. In a panel of 102 brands tracked across five assistants' APIs, 2.9% of citations pointed to the tracked brand's own domain, and 75.2% to other companies' pages, such as competitors, peers, and vendors in the same category (opens in a new tab) (preprint; March to May 2026; the author co-founded the tracking platform studied).

Being known is not the same as being recommended. Among 112 startups from the 2025 Product Hunt leaderboard, two models recognized 99.4% and 94.3% of the products when asked about them by name, but named them in only 3.32% and 8.29% of answers to discovery-style questions (opens in a new tab) (preprint; master's thesis). A recommendation needs sources that connect the brand to the buyer's need.

Where assistants find mentions

Studies count sources in two ways: as a share of all citations, or as the share of answers in which a source appears. A site can appear in many answers yet make up a small share of citations. Both views, with each study's setting:

The decision-stage figures come from an agency study of about 1,000 prompts on pricing, comparisons, alternatives, and "best for" questions, with 57,095 citations from ChatGPT, Perplexity, Gemini, and Google's AI Overviews and AI Mode, collected January 29 to February 4, 2026 (opens in a new tab). It found that third-party sources supplied about 80% to 95% of citations, depending on the industry (opens in a new tab).

Engines also differ. In API tests with ranking-style prompts, earned sources made up 93.5% to 95.1% of the domains ChatGPT cited, 86.3% to 87.3% for Claude, 67.4% to 73.4% for Perplexity, and 63.4% to 66.4% for Gemini, which cited brand sites most often (opens in a new tab) (preprint; mid-2025). The engine guides cover these differences, for example how to rank in Gemini and how to rank in Microsoft Copilot.

Two AI companies have data agreements with Reddit. OpenAI accesses Reddit's Data API to "better understand and showcase Reddit content, especially on recent topics" (opens in a new tab) (May 2024), and Google uses it to "better understand Reddit content and display, train on, and otherwise use it" (opens in a new tab) (February 2024).

What correlates with being named

• Mentions more than links. Across 75,000 brands, branded web mentions correlated with AI visibility at 0.664 in ChatGPT, 0.709 in AI Mode, and 0.656 in AI Overviews, and YouTube mentions at about 0.737; link metrics such as the number of backlinks showed "very weak correlations," and the number of site pages about 0.194 (opens in a new tab) (vendor study, December 2025; Spearman correlations). The study covered only domains with an Ahrefs Domain Rating above 40 (opens in a new tab), so it describes established brands.
• Branded links in between. In the same study, branded anchors correlated at 0.628 in AI Mode and 0.511 to 0.527 in ChatGPT and AI Overviews (opens in a new tab).
• Search rankings. On 10,000 finance and SaaS questions sent to the GPT-4o API, brands ranking on page 1 of Google showed a correlation of about 0.65 with mentions, Bing rankings about 0.5 to 0.6, and backlinks a weak or neutral relationship (opens in a new tab) (agency study, January 2025; statistical method not published).
• Referring sites and communities for new brands. For the 112 startups above, visibility in Perplexity correlated with referring domains (r = 0.319) and Reddit presence (r = 0.395), while a score of on-page optimization showed no correlation (opens in a new tab).
• Size. On a first answer to a generic prompt, household brands appeared 73% of the time, mid-market brands 44%, and small brands 11% (opens in a new tab).

What other sites say about an offer can outweigh familiarity. In laboratory tests of skincare recommendations, a well-known brand won every choice when products were equal, but a fictional competitor won half the time with a 0.075-star rating edge, 1.6 times the reviews, or a 7.3% lower price (opens in a new tab) (preprint). Ratings, review counts, and prices published on other sites are part of what an assistant can read about a brand.

What the evidence does not show

• Cause. The field studies above are correlational. A large brand has more mentions, more links, more searches, and more AI visibility at once, so the correlations cannot show that adding mentions raises recommendations.
• A proven technique. A review of 45 studies found no technique with a stable, long-term, cross-platform causal effect on discoverability (opens in a new tab) (preprint).
• Timing. No study has measured how long new mentions take to change AI answers. In the 102-brand panel, the fastest-moving small brands gained only 10 to 20 percentage points of visibility between March and May 2026 (opens in a new tab).
• Which mentions count most. The studies count mentions by source type and volume. None tests whether one mention on a widely cited page outweighs many on pages that are rarely cited.

What Google, Bing, and the FTC say

Google. Its guide to generative AI features in Search says those features "can show what's being said about products and services across the web, including in blogs, videos, and forum discussions. However, seeking inauthentic 'mentions' across the web isn't as helpful as it might seem." (opens in a new tab) Google's quality raters, whose ratings are used to measure how well Google Search works, are told to research reputation using independent reviews, references, expert recommendations, news articles, and other credible sources, and to be skeptical of claims websites make about themselves (opens in a new tab). Google says no single rating can directly change how a page appears in Search (opens in a new tab).

Bing. Bing judges a site's reputation partly by asking "What types of other websites link to the site?" (opens in a new tab) Its guidelines treat link buying and "artificial social promotion schemes that simulate popularity" (opens in a new tab) as abuse.

The U.S. Federal Trade Commission. Its rule on consumer reviews and testimonials, in effect since October 21, 2024 (opens in a new tab), prohibits (U.S. law):

• fake reviews, including "AI-generated fake reviews," and buying such reviews (opens in a new tab);
• incentives conditioned on a review expressing a particular sentiment (opens in a new tab);
• certain reviews by company insiders that do not disclose their connection (opens in a new tab);
• presenting a website the company controls as an independent source of reviews in its own category (opens in a new tab);
• buying or selling fake social media indicators, such as bot followers or views (opens in a new tab).

FTC staff say the rule does not ban incentives for reviews that carry no sentiment condition, but failing to disclose incentives could violate the FTC Act, as could failing to disclose relationships between influencers and brands (opens in a new tab). In December 2025, FTC staff warned 10 companies about possible violations and noted civil penalties of up to $53,088 per violation (opens in a new tab).

What buyers do after a mention

A mention starts a check. Among U.S. B2B professionals who use AI at work, when AI mentions a vendor, 71% visit the vendor's website, 63% search for the company on Google, 46% compare it with alternatives, and 38% check reviews on G2 or similar platforms (opens in a new tab) (vendor survey; 519 respondents; March–April 2026). The same buyers said they notice a vendor in an answer when it closely matches their use case (53%) or the description is clear and detailed (50%); 7% said brand recognition influences whether they notice a vendor (opens in a new tab). In a survey by a review platform, 45% of B2B software buyers said citations from software review sites are the most confidence-inspiring signal in an AI answer, and 64% said they encounter inaccurate AI chatbot recommendations often or very often (opens in a new tab) (1,076 respondents; March 2026; G2 runs a software review site).

Steps

These steps are AEO HQ's recommendations based on the evidence above; the facts in them link to their sources.

1. Find where you are mentioned now. Search for your brand while excluding your own site; Google's rater guidelines describe searching for reviews of a site this way, excluding pages on the site itself (opens in a new tab), as in "example.com" reviews -site:example.com. Add YouTube, Reddit, the review sites in your category, and directories. Record each page and what it says about you.
2. Find the pages AI assistants cite for your buyers' questions. Ask each assistant your buyers' unbranded questions several times and record the cited URLs. For Copilot, Bing Webmaster Tools reports the grounding queries behind your own citations and your Citation Share (opens in a new tab).
3. Correct the facts at the source. Update every profile you control with the same name, description, prices, and scope, and ask publishers to fix errors about you. The entity checklist lists the places to check.
4. Ask to be evaluated where buyers' questions are answered. Offer the publishers of cited comparisons and lists the facts they need to assess you: pricing, scope, customers you may name, and evidence. Disclose any relationship. Do not pay for placement in a list that presents itself as independent.
5. Ask every customer for a review on the platforms your buyers use. Do not tie an incentive to a positive review, and disclose any incentive. Do not ask only the customers you expect to be happy: FTC staff say that practice is not banned by the rule but "could violate the FTC Act" (opens in a new tab).
6. Publish original evidence others will cite. A dataset, a survey, or a sourced price comparison gives journalists and list authors a reason to name you.
7. Appear where buyers learn. Video and podcasts give other people's channels a reason to mention you; YouTube mentions had the strongest correlation with AI visibility in the 75,000-brand study (opens in a new tab).
8. Take part in communities under your own name. Answer questions where your buyers ask them, state your affiliation, and follow each community's rules.
9. Measure mentions and recommendations over time. Report the share of repeated runs that name you and your share of voice against named competitors, because lists of recommended brands change from run to run (opens in a new tab). Ask new customers where they heard of you. The method is in how to measure AI visibility.

What the evidence shows and does not show

Antipatterns

An antipattern is a practice that looks helpful but fails or backfires.

1. Buying mentions or list placements without disclosure. Google says seeking inauthentic mentions "isn't as helpful as it might seem" (opens in a new tab), and FTC staff say undisclosed relationships could violate the FTC Act (opens in a new tab). Instead, earn coverage and disclose any relationship.
2. Fake reviews, or incentives for positive ones. The FTC's rule prohibits both (opens in a new tab), with civil penalties of up to $53,088 per violation (opens in a new tab). Instead, ask every customer and disclose incentives.
3. Self-ranking "best" lists. In one test of AI answers about agencies, almost every frequently cited source was an agency's own list ranking itself first, and ChatGPT flagged the self-ranking in several answers (opens in a new tab) (48 answers; agency-published). A practitioner linked such lists to visibility losses at several brands in early 2026 (opens in a new tab). Instead, publish comparisons with a stated method, or leave ranking to independent sites.
4. Posting about your own company without saying who you are. The FTC's rule prohibits certain reviews by company insiders that do not disclose their connection (opens in a new tab), and Bing treats artificial promotion that simulates popularity (opens in a new tab) as abuse. Instead, participate under your own name.
5. Chasing link counts. Backlink counts showed very weak correlations with AI visibility (opens in a new tab), and Bing treats link buying (opens in a new tab) as abuse. Instead, seek mentions on pages buyers and assistants actually read.
6. Counting mentions without checking them. 64% of surveyed software buyers said they meet inaccurate AI recommendations often or very often (opens in a new tab). Instead, check what each mention says and correct wrong facts at the source.
7. Reporting "#1 in ChatGPT." Brand lists rarely repeat between runs (opens in a new tab). Instead, report rates over repeated runs.

Checklist

FAQ

What is a brand mention?

A brand mention is any page, video, review, post, or listing that names a company or product. It can include a link to the brand's site or not. For AI answers, the useful mentions are mostly on other sites, since only 2.9% of citations in one large study pointed to the brand's own domain (opens in a new tab).

Do unlinked brand mentions help SEO?

Neither Google's nor Bing's documentation reviewed for this guide describes unlinked mentions as a ranking factor. Bing's reputation test looks at the kinds of sites that link to a site (opens in a new tab). For AI answers, unlinked mentions still matter as text: Google says its generative AI features can show what's being said about products and services across the web (opens in a new tab), and web mentions correlated with AI visibility far more than backlinks did (opens in a new tab).

How do I get my brand mentioned by AI assistants?

Get your brand onto the pages assistants cite for your buyers' questions, and make sure those pages describe you correctly. That usually means independent comparisons and lists, reviews, video, and community discussion, plus a site that states your facts plainly. For ChatGPT specifically, see how to get recommended by ChatGPT.

How do I track brand mentions in AI answers?

Run a fixed set of buyer questions on each assistant several times over several weeks, and report the share of runs that name you. Bing Webmaster Tools reports citations and Citation Share for Copilot and Bing's AI summaries (opens in a new tab). See how to measure AI visibility.

Can I pay for mentions?

Not in ways that hide the payment. Paying for reviews that must be positive is prohibited by the FTC's rule (opens in a new tab), and FTC staff say undisclosed relationships between brands and influencers could violate the FTC Act (opens in a new tab). Google adds that inauthentic mentions are not as helpful as they might seem (opens in a new tab).

Does Reddit matter for AI recommendations?

It depends on the measure. Reddit appeared in about 62% of decision-stage answers in one agency study (opens in a new tab), but made up 3.3% of all citations in another panel (opens in a new tab). OpenAI (opens in a new tab) and Google (opens in a new tab) both have access to Reddit's Data API.

How long does it take for mentions to change AI answers?

No study has measured it. In one panel, the fastest-moving small brands gained 10 to 20 percentage points of visibility over about three months (opens in a new tab).

Change log

• September 28, 2026: First published.

Next steps

AEO HQ's Instant AEO Audit ($499) includes how AI assistants answer your buyers' questions, a side-by-side analysis of up to three competitors, a live crawl of your key pages, and AI crawler access checks.