What the AI Visibility Index Measured
Semrush pulled 126 million U.S. AI search prompts from January through April 2026 across four surfaces: ChatGPT, Gemini, Google AI Mode, and Google AI Overviews, then benchmarked brand visibility across 22 industries (Semrush). The first version of this Index, launched in late 2025, ran on 2,500 sampled prompts. A 50,000x jump in sample size in under a year is not a marketing detail. It is a signal that even the analytics vendors building AI visibility tooling are still discovering how unstable single-prompt measurement can be, and why category-level, high-volume sampling is the only way to produce a benchmark brands can act on.
The headline table below pulls the figures that matter most for a brand or marketing team deciding what to prioritize next.
| Metric | Finding |
|---|---|
| Marketing leaders who can fully track AI visibility | 9% |
| Marketing leaders who cannot measure AI visibility at all | 45% |
| Brands with top-100 visibility across all 4 platforms | 36 |
| Visibility concentration, Finance (top 3 brands) | 41.4% |
| Visibility concentration, News and Media (top 3 brands) | 82.9% |
| AI traffic growth to retail sites, Oct 2024 to May 2026 | +1,324% |
| AI traffic growth to travel sites, same period | +2,215% |
| Traffic or lead gains reported by integrated SEO+AI teams | 81% |
| Traffic or lead gains reported by siloed SEO+AI teams | 36% |
Mentions and Citations Are Still Not the Same Metric
The Index confirms a pattern our own measurement work keeps surfacing: being named by an AI assistant and being cited as its source are separate events, and most tracking setups only catch one of them. ChatGPT cites an average of 15 sources per response, while Gemini cites an average of three, and on Gemini the overlap between which brands get mentioned and which domains get cited was as low as 30% (Semrush). A brand can be the answer without owning the footnote, and it can own the footnote without being named in the text a user reads.
This is not a new problem for Semrush's research. An earlier Semrush study of 3,981 domain appearances found that 62% of AI citations never produce a brand mention at all, and that ChatGPT and Gemini behave almost inversely: ChatGPT cites 87% of the time but names the brand in only 20.7% of answers, while Gemini names brands 83.7% of the time but cites the source only 21.4% of the time (Semrush). What the 2026 Index adds is scale and an industry lens: the mention-citation gap is not an edge case in a small sample, it holds across 126 million prompts and 22 categories. We break down what that means for measurement setup in our guide to tracking brand mentions in AI search.
The Real Headline Is the Measurement Gap
Visibility rankings get the press coverage, but the 45%-cannot-measure and 9%-have-full-coverage figures are the more consequential finding for most marketing teams. Most AI visibility tooling in market today was built on rank-tracking logic: a keyword, a position, a URL. AI answers do not produce a stable position to track. The same prompt returns a different brand set depending on the model version, the user's location, prior conversation context, and the day it was asked. A team using single-platform, single-prompt monitoring is not measuring visibility. It is sampling noise and calling it a trend line.
That gap is also where competitive advantage sits right now. If 91% of marketing leaders lack full-platform tracking, then a brand that closes even part of that gap gains a view into its AI presence that most of its category cannot see. That is the practical case for a dedicated AI visibility audit: not to produce a vanity score, but to establish the baseline that 91% of teams, per this data, currently do not have.
Why Visibility Concentration Differs So Much by Industry
The 41.4% to 82.9% range in top-three brand concentration is a category-strategy signal, not a footnote. In News and Media, where 82.9% of visibility sits with three brands, the incumbents' scale, publishing cadence, and existing citation history make them difficult for a challenger to dislodge from AI answers in the near term. In Finance, where the top three hold only 41.4%, the field is far more open: no brand has locked in the kind of dominance that AI models default to when a query is ambiguous.
The practical implication is that a generic "get cited more" strategy is the wrong frame. A finance brand entering this data should be building toward category leadership, because the ceiling is genuinely reachable. A brand in a concentrated vertical like News and Media should instead target specific sub-topics or query types where the incumbents are thin, rather than trying to out-publish them on volume.
Integrated Strategy Is Outperforming Siloed Strategy
The 81% versus 36% split between brands with integrated SEO and AI visibility strategies and those managing the two separately is the Index's clearest actionable finding. The mechanism is straightforward once you look at what earns an AI citation: structured, entity-clear content with third-party corroboration performs well in both organic rank and AI retrieval, because the underlying signals overlap. Teams that treat AI visibility as a bolt-on project, run by a different owner with a different content calendar, end up duplicating work or optimizing for a channel that no longer reflects where their traffic originates.
This mirrors what we've seen building visibility programs for clients: the brands that treat AI visibility as an extension of their existing search strategy, rather than a parallel initiative, move faster because they are not re-litigating content decisions twice. Our framework for building that kind of unified approach is in our guide to building an AI search visibility strategy.
What Brands Should Do With This Data
- Set up tracking across all four platforms the Index measured (ChatGPT, Gemini, Google AI Mode, Google AI Overviews), not only the one your team checks manually.
- Separate mention tracking from citation tracking in your reporting. Treat them as two metrics with two different remediation paths, not one combined "AI visibility" number.
- Benchmark your category's concentration level before setting visibility targets. A 10-point visibility gain means something different in Finance than in News and Media.
- Move AI visibility ownership into the same team and workflow as SEO. The 81% vs. 36% gap suggests structural integration matters more than tool selection.
- Re-baseline quarterly. A sample built on prompts from January through April will not describe your visibility in Q4, especially given how fast model versions and citation behavior are moving.
If you want a current read on where your brand sits against this data, Elaventra's free AI Visibility Report benchmarks your mention and citation rates across the same four platforms the Index covers, and our AI Visibility Strategy Call walks through what to prioritize based on your category's concentration level.
Frequently asked questions
What is the Semrush AI Visibility Index?
It is a benchmark study, expanded in June 2026, that analyzed 126 million U.S. AI search prompts from January through April 2026 across ChatGPT, Gemini, Google AI Mode, and Google AI Overviews to measure brand visibility across 22 industries.
What is the difference between an AI brand mention and an AI citation?
A mention is when an AI assistant names your brand in the text of its answer. A citation is when it links to or footnotes your content as a source. Semrush's research shows these overlap far less than most teams assume: on Gemini, brand mentions and cited domains overlapped by as little as 30%, and a separate Semrush study found 62% of citations produce no brand mention at all.
Which industries have the most concentrated AI visibility?
News and Media is the most concentrated category in the Index, with the top three brands holding 82.9% of category visibility. Finance is among the least concentrated, with the top three brands holding 41.4%, leaving more room for challenger brands to gain ground.
How often should brands measure AI visibility?
Given how fast model versions and citation patterns shift, quarterly re-measurement is a reasonable minimum, and monthly tracking is better for categories with active competitive movement. A single point-in-time check will not hold up as a strategy baseline.
The Takeaway for Brand and Marketing Teams
The most useful number in Semrush's 2026 AI Visibility Index is not a ranking. It is the admission, from 126 million prompts of evidence, that 91% of marketing leaders are operating without full visibility into how AI assistants represent their brand. Rankings will keep shifting as models update. The teams that build real measurement now, across platforms and across the mention-citation split, are the ones who will know it when the rankings move.