Being cited by an AI search engine is widely seen as an authority signal — proof that your content is credible enough for the AI to source from.
But being cited is not the same as being mentioned. And in more than half of cases, a citation doesn’t instantly lead to a brand mention.
Kevin Indig coined the term ghost citation to describe this gap. When an AI cites your site without explicitly mentioning you, readers walk away without recognizing your brand or trusting you as the original source.
To investigate this, we partnered with Kevin and analyzed thousands of domain appearances across multiple AI search engines.
Key takeaways
- 62% of AI citations are ghost citations. Your site gets a source link, but the AI never says your name in the answer.
- Each AI engine behaves differently. When a brand appears in a Gemini answer, it's named in the text 83.7% of the time but cited as a source only 21.4% of the time. ChatGPT does the reverse: 87% citation rate, 20.7% mention rate.
- Mention rates vary by country. Brands get mentioned in 50% of AI answers in India and Sweden, but they only get mentioned in 18%-22% of AI answers in Italy, Brazil, and the Netherlands.
- Short, conversational queries produce 30x-50x more brand mentions. Long prompts trigger more citations, but they lead to fewer brand mentions.
- Query intent and content type strongly affect brand mention rates. Informational content earns citations but rarely brand mentions. Comparative content produces 2.4x more brand mentions.
Methodology
We ran this study in collaboration with Kevin Indig and Growth Memo using data from the Semrush AI Visibility Toolkit.
We logged 3,981 domain appearances for 115 prompts run across 14 countries and four leading AI search engines:
- ChatGPT
- Google AI Overviews
- Gemini
- Google AI Mode
For every domain appearance, we noted two outcomes:
- Cited: The domain appeared as a source link in the response
- Mentioned: The brand name appeared in the AI answer
We also calculated the mention rate across different query types.
Let’s explore the findings in greater depth.
1. 62% of citations don’t lead to brand mentions in AI answers
You might assume being cited means your brand is visible in AI answers. The data says otherwise.
Every appearance in our dataset fell into one of three buckets:
- Almost 62% (61.7%) were ghost citations. AI platforms used the page as a source link, but the brand name never appeared in the actual answer.
- Over 13% (13.2%) were both cited and mentioned. The source link plus brand name appeared in the answer.
- Only about 25% (25.1%) were brand mentions without a citation. The AI named the brand in the answer without linking to a source page.
That means 74.9% of all brand appearances included a citation, but only 38.3% of appearances included a brand mention. So, the citation rate is nearly double the mention rate.
Think about it this way: Appearing as a source in AI search doesn’t automatically make your brand visible to users. A citation offers attribution, while your brand stays absent from the answer itself.
2. Every AI engine behaves differently
The four AI engines we analyzed treat citations and mentions in fundamentally different ways:
- Gemini mentions brands in 83.7% of appearances, but it only generates a citation link 21.4% of the time. An “appearance” here means a domain showed up in an AI answer at all, whether as a source link, a brand mention in the answer text, or both. So, Gemini acts like a conversationalist, drawing on what it already knows.
- ChatGPT does the opposite of Gemini. It cites brands 87% of the time but mentions brands in only 20.7% of answers. These answers look more like academic papers with footnotes.
- Google AI Overviews sit in the middle, but they lean toward citations.
- Google AI Mode mentions brands at nearly twice the rate of ChatGPT,but AI Mode still acts closer to a footnoted research piece than to Gemini’s knowledge-based answers.
This means you can’t assume visibility in one AI engine will translate to visibility in another. In our dataset, there was almost no overlap between the brands ChatGPT cited and the brands Gemini named for the same prompts. Different engines reward different signals, formats, and sources.
This means you can’t assume visibility in one AI engine will translate to visibility in another. Different engines reward different signals, formats, and sources.
“There’s almost no overlap between which brands ChatGPT cites and which ones Gemini names for the same prompt. These are different behavioral systems. Treat them that way.”
3. Strong brands get named, while aggregators get cited
Among the domains that appeared most often in our data, aggregator and academic sites like medium.com, wikipedia.org, wired.com, and harvard.edu tended to be cited but not named.
Brands with strong public identities were showing up the opposite way: They were named in the answer, often without a citation.
For example, Google was named in AI answers nearly three times more often than it appeared as a source link. Apple's brand name appeared in answers nearly twice as often as it was cited.
At the other end of the spectrum, medium.com was cited 16 times in our dataset and never named in any AI answer.
The chart values are normalized to assume 100 citations. A value of 100 means a brand was mentioned and cited equally. Google’s 278 means Google was named 2.78 times for every 100 citations of a Google URL. Medium’s 0 means medium.com was cited but never named in the answer text.
As Kevin puts it:
“The AI knows the information about the brand came from somewhere, but doesn’t feel the need to explicitly say so to users. The brand name carries on its own.”
If you’re a publisher or research-style site, expect more citations than mentions and track citations as your primary AI visibility metric.
If you’re a consumer brand, aim for the reverse: prioritize tracking unlinked brand mentions on third-party websites (Reddit threads, news coverage, listicles) alongside your backlinks. Both feed AI engines’ familiarity with your brand.
AI engines mention brands they’ve already seen consistently across the web. PR, community discussions, and clear brand language all reinforce that familiarity.
4. AI engines disagree on the same brand 22% of the time
When running this analysis, we tested 454 prompt-and-domain combinations across multiple engines. Each combination is one specific prompt paired with one specific domain. For example, the prompt “best CRM for small business” paired with hubspot.com, tested across each AI engine.
In 100 of the 454 (22%), the engines disagreed on whether to name the brand.
For example:
- Instagram.com was mentioned by ChatGPT and Gemini. Google AI Overviews cited it without naming it.
- Facebook.com was named in three out of three Gemini appearances. Google AI cited Facebook nine out of nine times and named it once.
For accurate measurement, check each engine separately for your top branded and non-branded topics. Aggregated scores might smooth over the cases where you’re visible in one engine and invisible in another.
5. AI brand mention rates vary by country
Brand visibility patterns in our dataset differed sharply by market. The combination of citations and brand mentions varied especially.
For example:
- India and Sweden: 50% of AI answers to queries from these markets included a brand name in the response. That's the highest in the study, and query patterns there leaned more conversational and brand-forward.
- Italy, Brazil, and the Netherlands: Only 18%-22% of AI answers included a brand name, even though AI engines cited sources from these markets at very high rates (82%-94%).
- The U.K. and Canada (41% and 44%) sit in the upper-middle of the range. Brands are mentioned in just under half of AI answers from these markets.
If you operate across multiple markets, expect AI visibility patterns to look different in each. Set country-specific goals for citations and mentions rather than applying a single playbook globally.
Note: We localized and verified all prompts for each market, so the differences in mention rates reflect genuine regional patterns.
6. Short, conversational queries produce 30x-50x more brand mentions
Query phrasing changes brand mention outcomes dramatically. Across the prompts in our dataset, the length ranged from 16 to 98 characters, averaging about 60.
Short, conversational queries have brand mention rates of nearly 100%, while long, structured prompts have mention rates of just 2%-3%. That’s a 30x-50x difference on the same topic.
A short query like “Should I lease or buy a car for my business?” leads to brand mentions nearly every time. A long version of the same question with extra context and framing (e.g., ), produces almost no brand mentions even though it triggers more citations.
“Brands need to map not just which topics they want to appear in, but which phrasing patterns produce mentions versus ghost citations. Short, conversational queries and long, structured queries behave like different products.”
This means brand visibility in AI search is not just topic-dependent. It is also highly sensitive to how users ask the question.
7. Informational queries trigger the most ghost citations
Content type correlated strongly with brand mention rates in our data.
Informational content earned high citation rates but low mention rates, while comparative content earned mentions more often:
- Informational queries (“what is”, “explain”, “how does”): 89.3% citation rate, 18% mention rate
- Comparative queries (“best”, “vs”, “recommend”): 43.3% mention rate, 2.4x more brand mentions than informational
- How-to queries: 42.8% mention rate
- Commercial queries (pricing, buying intent): 35.6% mention rate, 84.4% citation rate
The pattern makes a lot of sense: Informational content gets consumed as raw reference data. Comparative content forces the AI to name the players it’s evaluating.
How to close the citation gap
Citations and mentions are separate outcomes.
Citations reflect domain authority and require original content and topic authority. While mentions reflect brand trust and come from clear positioning, positive sentiment, niche authority, and thorough product content.
To close the gap, develop strategies to get both cited and mentioned, analyze citations across LLMs and different categories, and identify low-hanging opportunities.
In Semrush’s AI Visibility Toolkit, you can track your AI search visibility across key areas:
- By AI engine. Citation and mention rates differed sharply by LLM in our data. Analyze patterns and identify opportunities for each AI platform.
- By country. Mention rates ranged from 18% to 50% across markets. Set country-specific goals based on your business priorities.
- By intent. Informational, comparative, how-to, and commercial queries produced different mention rates. See which topic areas you perform well in and where you’re missing visibility in AI search.
You can also uncover opportunities to increase brand visibility from content that AI platforms already cite.
Open the Visibility Overview report and scroll to “Your Performing Topics.” Here, you can identify topics and prompts where your content is cited, but your brand isn’t mentioned.
This means the page references your insights without clearly naming your brand or products.
In some cases, adding clearer brand references, product mentions, or comparison sections featuring specific tools and companies can help close that gap.
