Semrush study finds most AI citations are ghost citations
Semrush found 61.7% of AI citations were ghost citations, where a source was linked but never named. That can flatter visibility dashboards while brand recall stays weak.
Semrush’s latest AI visibility work lands a blunt warning for agencies selling “citation” wins: being linked in an answer is not the same as being remembered in it. The study, built with Kevin Indig, found that 61.7% of AI citations were ghost citations, meaning the system pointed to a domain but never named the brand in the response.
That gap matters because a source link can preserve authority without delivering recognition. Semrush analyzed 3,981 domain appearances across 115 prompts in 14 countries and four systems, ChatGPT, Google AI Overviews, Gemini and Google AI Mode. Only 13.2% of appearances were both cited and mentioned, while 25.1% were mentions without a citation. In other words, a clean source report can still hide the fact that the brand never entered the user’s head.
The platform split was sharp. Gemini named brands in the text much more often than it cited them, while ChatGPT skewed hard toward citation without mention. Geography also changed the picture: India and Sweden showed higher mention rates, while parts of Europe and Latin America lagged. Query shape mattered too. Short conversational prompts produced 30 to 50 times more brand mentions than longer prompts, and comparative queries generated 2.4 times more mentions than informational ones. That makes prompt design and intent coverage part of the visibility game, not just a side note.
For SEO agencies, the measurement problem is now obvious. Reporting that stops at source inclusion can inflate the appearance of success while the client’s brand remains invisible in the answer itself. Semrush’s earlier AI visibility research already showed Wikipedia and Reddit outranking corporate websites in AI citations, which only sharpens the point: authority alone is no longer enough if the model is elevating third-party ecosystems and comparison pages over branded language.
.png&w=1920&q=75)
The broader search shift explains why this is happening. Google launched AI Overviews to all U.S. users on May 14, 2024, expanded them to more than 100 countries in October 2024, then said they were available in more than 200 countries and territories and more than 40 languages after a May 2025 rollout. Google also said the inline links added in October 2024 sent more traffic to supporting websites than the earlier design. AI Mode followed in March 2025, pushing search deeper into follow-up questions and multi-turn exploration. OpenAI’s May 2024 partnership with Reddit, built around access to Reddit’s Data API for real-time, structured, unique content, reinforced how much community and comparison sources now shape answers.
The practical takeaway is simple and uncomfortable for agencies charging premium retainers: rewrite the scoreboard. Track mentions separately from citations. Build campaigns that force brand language into comparison content, review ecosystems and trusted third-party sources. In AI search, a link without a name is a weak win, and weak wins do not protect pricing power for long.
This article was produced by Prism’s automated news system from verified source data, official records, and press releases, then run through automated quality and moderation checks before publishing. The system is built and supervised by the people who set the standards it runs under. Read our full AI policy.
Did this article answer your question?
