In short
AI Overviews are the AI-generated answer blocks that Google has been displaying above the classic list of results since 2024. Instead of just listing links, Google uses its language model Gemini to summarise the answer to a search query in a few sentences and links to individual sources.
For users this means that many questions are answered directly on the search results page without a click on a website being necessary. For website operators it means a fundamental shift: search is turning from a list of links into an answer engine, and visibility increasingly means being cited in this AI answer rather than just ranking first. AI Overviews are thus the most visible expression of what Google itself describes as the shift to "generative search".
From SGE to AI Overviews and AI Mode
The feature started in 2023 as an experiment called Search Generative Experience (SGE) in Google Labs. At Google I/O in May 2024 it was moved into regular search as AI Overviews. In Germany, AI Overviews have been rolled out since 26 March 2025, initially in German and English. A distinction has to be made with AI Mode, a separate, dialogue-oriented search mode that has also been available in Europe and Germany since October 2025 and conducts search more like a chat. AI Overviews are therefore the answer block within classic search, while AI Mode is a separate, consistently conversational interface.
How AI Overviews work technically
Behind an AI Overview is a process that Google describes as "query fan-out": a search query is broken down into several sub-questions, these are answered in parallel and the results are condensed into a coherent answer. The model draws on content from the Google index and links selected sources as evidence. For a page to be considered as a source at all, its content has to be accessible to the machine: rendered on the server, clearly structured and thematically unambiguous. This is exactly where generative engine optimisation (GEO) comes in: optimising content so that it is cited in AI answers.
Table: classic search vs AI Overviews
| Aspect | Classic list of results | AI Overviews |
|---|---|---|
| Presentation | ten blue links | AI summary with sources |
| User action | click on a result | answer often without a click |
| Visibility goal | ranking position | citation in the answer |
| Basis | relevance and link signals | Gemini + Google index |
| Consequence for traffic | measurable click | more zero-click searches |
What AI Overviews mean for SEO
The most important consequence is the rise in zero-click searches: when the answer is already at the top, the click-through rate to the organic results below falls, especially for purely informational search queries. At the same time, a new form of visibility emerges: being mentioned as a cited source in the AI Overview. Anyone who appears there gains reach and authority, even without the classic click. SEO is thus shifting from pure ranking optimisation towards the question of how content becomes worth citing: is it unambiguous, up to date, well structured and backed by reliable facts? How companies approach this in concrete terms is described in our article AI visibility: can AI even find your website?.
Which search queries trigger AI Overviews
Not every search gets an AI Overview. Google mainly displays them when a question benefits from a summarising answer, typically for informational and explanatory queries ("how does ... work", "what is the difference between", "why"). For clear navigational searches (a specific brand or website) and many transactional queries they appear less often, because the user wants to click through in a targeted way anyway. In practice this means that guide, definition and comparison content is most affected by the shift, while product, brand and local searches remain more robust. Anyone who knows their own content mix can estimate which part of the traffic will come under pressure, and where investing in citable content pays off most.
How to make content citable
For a page to appear as a source in an AI Overview, it has to be easy for the machine to capture and cite. Four levers have proven effective. First, technical accessibility: server-side rendering, a clean structure and valid structured markup, so that the crawler sees the content in full at all. Second, a clear answer structure: one concrete question per section, answered in the first sentences and followed by more depth. Third, verifiability: checkable facts, figures with sources, up-to-dateness; models prefer content that can be verified. Fourth, thematic clarity: a page that covers one topic exhaustively is more likely to be cited than one that touches on many topics. These principles overlap strongly with good classic SEO; the difference lies in the goal: not the click, but the mention in the answer. Anyone who does not want to implement the four levers themselves can get them from our agency for generative engine optimisation as a package of audit, technology and content.
A practical example
A mid-sized company sells height-adjustable desks and has ranked on page one for "ergonomic workplace tips" for years. With the introduction of AI Overviews, Google now answers the question directly with a list of recommendations, citing three sources, and the desk retailer is not among them. The click-through rate to its guide drops. An analysis shows that the content is good, but it does not answer the question in a clearly citable form. After a revision (precise definitions, a concise step-by-step list, clear subheadings), the page appears as a cited source in the AI Overview. Traffic via the classic click remains lower than before, but qualified visibility via the AI answer is added.
Measuring the impact
Because AI Overviews come between search and click, they change the figures in Google Search Console in a characteristic way: the impressions for a page can remain stable or rise, while the clicks, and with them the click-through rate, fall, an indication that the answer is increasingly taking place in the overview. What is revealing is therefore less the absolute number of clicks than the development of the click-through rate per type of search query over time. It is also worth checking whether your own domain appears as a source in AI Overviews, either manually through samples of relevant search queries or with specialised monitoring tools that capture this systematically. It is important to adjust expectations: a decline in the click-through rate for informational queries is not necessarily a quality problem of the page but a reflection of the changed search interface. Sensible goals shift accordingly, from pure click maximisation towards brand presence, citation frequency and qualified conversions. Whether your website meets the technical requirements for being cited is shown by a free GEO check in 30 seconds.
Frequently asked questions about AI Overviews
Are AI Overviews and AI Mode the same thing?
No. AI Overviews are the AI answer block within classic Google search. AI Mode is a separate, consistently dialogue-oriented search mode that treats queries like a chat. Both use Google's Gemini models but differ in presentation and use.
Do AI Overviews harm my website traffic?
For purely informational search queries, click-through rates often fall because the answer is already at the top. At the same time, being cited as a source creates a new kind of visibility. Transactional and brand searches are less affected.
How do I get into an AI Overview?
By making content machine-readable, unambiguous and worth citing: rendered on the server, clearly structured, factually reliable and focused on a concrete question. This approach is summarised under the term GEO and is part of modern SEO and marketing work.
Source and background: Google on generative AI in search.