Google AI Overviews are AI-generated summaries that appear at the top of search results for certain queries, synthesising information from multiple web sources into a single answer with attributed citations. They are Google's most significant change to the search results page since the introduction of featured snippets.
An AI Overview is a boxed, multi-sentence answer that appears above organic search results. It is generated by a large language model that has read a selection of web pages and synthesised their content into a response, with small numbered citations linking to the source pages. Unlike featured snippets, which show a verbatim excerpt from one source, AI Overviews are synthesised: they may draw from three to eight sources, combine points from each, and produce prose that does not appear verbatim on any of the underlying pages.
As of mid-2026, AI Overviews appear on roughly 30 to 40 percent of queries in the UK, with higher rates on definitional and how-to queries and lower rates on transactional, navigational, and highly contested queries. Google is selective: queries where the answer is genuinely complex, disputed, or where a synthesised answer might mislead (financial advice, medical guidance) are less likely to generate an Overview. The feature is expanding, but it is not universal, and it is not replacing organic results for all query types.
This is the question every publisher needs to understand, because citation in an AI Overview is now a meaningful traffic and visibility lever in its own right. Google has not published a precise algorithm for citation selection, but the pattern is consistent enough to draw practical conclusions.
Topical authority matters more than domain authority. Specialist sites covering a narrow subject in depth are frequently cited ahead of high-domain-authority generalist sites. A specialist property law publication is more likely to be cited for a leasehold query than a national news outlet that covers property occasionally.
Specificity is consistently rewarded. Pages that contain a clear, direct answer to the query in the first paragraph, supported by specific data or a concrete example, are cited at higher rates than pages that take several paragraphs to reach the point. The model appears to prefer content that can be extracted cleanly and attributed accurately.
Structure helps. Pages with clear headings that map to sub-questions, definitions set off from surrounding prose, and numbered or bulleted steps for how-to content are more frequently cited than pages with the same information buried in long, undifferentiated paragraphs. The model reads structure as a signal of clarity.
E-E-A-T signals remain relevant. Pages with author bylines, publication dates, cited sources, and clear organisational attribution are cited more reliably than anonymous or undated content. This is especially pronounced in YMYL topics, but the pattern holds more broadly.
The most important shift is this: for informational queries, being cited in the AI Overview is now often more valuable than ranking first in organic results. The Overview appears above organic results and captures a large share of attention for queries where it appears. A site cited in the Overview with a visible link gets brand exposure and a meaningful proportion of clicks. A site that ranks first in organic results but is not cited in the Overview may find its click-through rate has fallen substantially from two years ago.
The practical implication is that content strategy for informational queries now needs to optimise for citation, not just ranking. Those two goals are largely aligned (the same signals help with both), but citation optimisation has some distinct requirements: answer-first structure, clear attribution, specific data, and a reputation for reliability on the topic that extends across multiple pages, not just the one being optimised.
For transactional and commercial queries, organic results remain relatively unchanged. AI Overviews are less common for "buy X" or "best X for Y" type queries, and where they do appear, they tend to link to product pages and aggregators rather than editorial content. Publishers in commercial categories should not assume the same shift applies to their primary query types.
AI Overviews create a real traffic headwind for sites whose primary value was aggregating information that is now freely synthesised by Google. If your content strategy is based on being the clearest explainer of widely available information, that value has been partially captured by the Overview itself. The content that retains its value is the content the model cannot synthesise: first-hand data, original research, direct product experience, and specialist analysis that does not appear anywhere else. That is the content that gets cited, and the content that drives clicks when cited.
Related: What is answer engine optimisation? · Is SEO dead in 2026? · What is topical authority in SEO?
Quarrybank structures your content for AI Overview citation from the brief stage, so your pages are the ones Google's model reaches for.
Request founding access