Semrush adds Gemini filtering to AI Visibility Toolkit as mentions outpace citations
Semrush has published a guide to tracking brand visibility in Google's Gemini app, alongside a Gemini filter in its AI Visibility Toolkit that separates mentions, citations and cited pages for the platform. The company's Ghost Citations study found Gemini names brands in 83.7% of their appearances but cites them in only 21.4%, the reverse of ChatGPT, which cites brands in 87% of appearances and mentions them in 20.7%.
The gap matters for measurement. In Semrush's own data for asana.com in the US database as of 28 October 2026, Gemini produced roughly 11 mentions for every citation, and Gemini accounted for 24% of Asana's AI mentions against ChatGPT's 15.4%. In a test of 20 buyer prompts, Asana appeared in 14 Gemini answers but was cited in one, and in one case Gemini named it the top choice while crediting a competitor's site.
The toolkit's Gemini view also lists topic opportunities, where competitors are mentioned but the tracked brand is not, and performing topics where the brand already appears. Semrush puts Asana at about 4.4K performing topics and 7.1K opportunities against its suggested competitor set, drawing on a database of more than 317 million prompts and responses across 117 regional databases.
For brands, the practical consequence is that judging Gemini by referral traffic undercounts its influence, since buyers who see a name in an answer may search for it later and arrive as direct or branded traffic. Semrush recommends tracking mentions as the primary Gemini metric and running manual prompt checks in signed-out, incognito sessions before scaling with the toolkit.
Reported by Semrush Blog ↗
Topics: Gemini, Measurement, Citations
In this directory: Semrush AI Visibility Toolkit