Semrush study: AI visibility metrics conflate mentions and citations, obscuring what actually needs fixing

Semrush's 2026 AI Visibility Index, covering 22 categories and four AI platforms, found that fewer than half of top-100 mentioned brands also ranked in the matching top-100 citation lists on average each month. The study shows that a single AI visibility metric obscures whether a brand appears in the answer itself, whether its own domain supplies evidence, or whether third-party sites dominate citations—three distinct problems requiring different fixes. Examples include Ford, which appeared in all top-100 mention lists in the U.S. but ranked outside the top 1,000 for citations across platforms, and Tesco and Asda, which both ranked in the top 100 for mentions on multiple platforms but diverged sharply on citation rankings. The analysis notes that citation patterns vary significantly across platforms: ChatGPT favors news and information sites, AI Overviews emphasizes streaming and business software, and Gemini differs from both in source mix and mention volume. Within-category trends also matter more than broad platform comparisons; real estate shows consistent recurring sources like Apartments.com and Zillow, while industrial searches lack an equivalent stable core. MARTECH recommends establishing separate mention and citation baselines for specific customer questions and inspecting cited pages before allocating resources to optimization.