Google AI Overviews now reach over 2 billion monthly users across more than 200 countries. That is roughly a quarter of everyone on Earth receiving AI-generated summaries before they even see a traditional search result. New analysis of the 50 most-cited domains in AI Overviews gives us the clearest picture yet of who is being recommended at that scale - and the structural reasons why.
If you are working on AI visibility for a brand, this kind of citation data is not just interesting. It is the closest thing we currently have to a map of how Google decides what to surface. Studying the winners tells you what the system is optimising for.
Institutional authority still dominates
The most-cited domains are not surprises. Wikipedia sits near the top, alongside government and academic sources, major news publishers, and established reference sites. These are organisations that have spent years - in some cases decades - building content that is accurate, well-structured, frequently updated, and heavily cited across the web.
What this tells us is that Google's AI citation model is heavily weighted towards sources it already has strong trust signals for. This is not a new content algorithm being applied from scratch. It is the existing authority model being fed into a generative layer. Brands hoping to appear in AI Overviews purely by producing more content will be disappointed if they do not also address the underlying credibility signals those top-cited sites have in abundance.
For UK businesses, this means the basics still matter: structured, well-sourced content; a clear organisational identity that Google can verify; and genuine third-party mentions from credible sources. None of that has changed. What has changed is how consequential it now is to get right.
Community and discussion platforms are holding their own
Reddit's continued presence in AI Overview citations is notable. It has become a go-to source for first-person experience content - product reviews, how-to advice, comparisons, troubleshooting. Google appears to weight this type of content specifically for queries where experiential insight matters more than encyclopaedic accuracy.
This has direct implications for brand strategy. Your customers are likely discussing your products, your category, and your competitors on Reddit and similar forums. If those discussions skew negative, or simply never mention your brand at all, AI Overviews will reflect that absence. Monitoring community platforms and actively creating reasons for people to discuss your brand positively in public forums is no longer a fringe tactic - it is part of AI visibility work.
What vertical coverage in the top 50 reveals
Looking at which sectors the top-cited domains represent gives a useful signal about where AI Overviews are most active. Health, finance, technology, and reference content feature heavily. These are query categories where users want concise, reliable answers - exactly the conditions under which AI Overviews are most likely to appear.
For brands operating in these verticals, the competitive implications are significant. If your category is one where AI Overviews frequently fire, and your brand is not among the sources being cited, you are effectively invisible at the top of the page for a substantial proportion of relevant queries. That is not a future risk. Given the scale of AI Overviews reported here, it is a present one.
Sectors not well-represented in the top 50 - local services, niche B2B, specialist trades - should not read this as irrelevance. AI Overviews still fire for those queries; the citations are just more distributed. The opportunity to be a dominant cited source in a narrower vertical is arguably more achievable than trying to compete with Wikipedia for broad informational queries.
Citation frequency is not the same as citation quality
Being cited 10,000 times across a range of low-stakes informational queries is a very different position to being cited consistently for high-intent, commercial queries in your specific sector. The aggregate citation counts in data like this can obscure that distinction. A niche brand cited reliably whenever someone asks about its specific product category is in a stronger commercial position than a general publisher cited constantly for tangential topics.
This is why AI visibility measurement needs to move beyond simply tracking whether your domain appears in AI Overviews at all. The queries for which you are cited matter enormously. Are you being cited in response to brand-name searches? Category comparisons? Buying intent queries? Or only broad informational questions where purchase decisions are far off? Mapping citation context to the purchase funnel is the next layer of sophistication for brands that have already started this work.
What the top-cited sites actually have in common
Strip away the brand names and the most-cited domains share a set of observable characteristics. Their content answers specific questions directly. It is updated regularly. It cites its own sources. It has clear authorship or organisational provenance. It is structured so that individual sections can be extracted and understood without surrounding context.
That last point is underappreciated. AI Overviews do not cite entire websites - they pull specific passages. Content that is written as a coherent whole, where meaning depends heavily on reading everything in sequence, is harder to extract than content built around clear, self-contained answers. Headers that reflect actual questions. Paragraphs that open with the key point. Definitions that do not assume prior reading. These are structural choices that directly affect citability.
For marketing teams, this is a practical brief. Review your most important content assets - not for keyword density or word count, but for extractability. Could Google's AI system pull a single paragraph from that page and have it make sense on its own? If not, the content may be doing fine for traditional SEO while actively underperforming for AI citation. Those are increasingly different optimisation targets, and the gap between them is worth closing deliberately.
The window to establish citation presence is now
Citation patterns in AI systems tend to reinforce themselves. Once a source is consistently cited, it accumulates more third-party references, more traffic, and more implicit trust signals - which in turn make it more likely to be cited again. The domains at the top of this list did not get there overnight, but many of them will be significantly harder to displace in two years than they are today.
For brands that have not yet invested seriously in AI visibility, that trajectory is the most important thing to take from this data. The question is not whether AI Overviews matter to your category. At two billion monthly users, they almost certainly do. The question is whether you are building citation credibility now, while the field is still being established, or waiting until the positions have hardened.