AI Overviews are changing how people get answers in Google Search. Instead of a list of links, a summarised answer generated from several sources sits at the top. For SEO this does not mean that old rules are suddenly wrong. What it is really about is a shift: visibility increasingly arises where your own brand enters an answer as a reliable source, even if nobody clicks afterwards.
For marketing this has tangible consequences. Many marketing departments still measure their success in search by clicks and rankings. If part of the questions is already answered on the results page, these numbers only tell part of the story. Anyone who does not understand this draws the wrong conclusions and invests in the wrong place. They may cut content that is valuable for the brand, only because it generates fewer visits than before.
- AI Overviews summarise answers directly in search and link to sources.
- According to Google, there are no additional technical requirements to appear there.
- The proven foundations of SEO remain the prerequisite for visibility.
- Clicks alone are no longer enough as a yardstick to judge impact in search.
- Brands win when their content is unambiguous, correct and well structured.
Table of contents
What AI Overviews really change for SEO
In its documentation on AI features in search and your website, Google describes very soberly what matters. There are no additional requirements to appear in AI Overviews or in AI Mode. A page must be indexed and generally eligible to be shown with a snippet in search. Special files or dedicated markup for AI are not needed.
I find this clarity refreshing. It takes the wind out of many promises that talk about secret tricks for AI search. What counts are the foundations: pages that may be crawled, sensible internal linking, important content in text form and structured data that matches the visible content.
So what has changed is less the technology than the effect. AI Overviews answer a question before anyone visits a website. The brand can be named as a source in the process. Whether a visit follows depends on whether the user wants to know more than the summary offers. For many simple questions the answer is enough, for questions with real consequences usually not.
For SEO this means: being in first place is no longer enough. That is why people increasingly talk about GEO instead of SEO, generative engine optimisation aimed at the answers of generative systems. I find the term useful as long as it stays clear that it builds on the same foundations. Your own content has to be clear enough to be quoted and deep enough that a click is still worth it. This also includes marking up person and brand unambiguously with structured data according to Schema.org, for example as Person and Organization, so that search systems understand who is behind a piece of content.
Why AI Overviews are harder to read in fashion
In fashion the situation is special. Many searches are not purely informational but combine curiosity, inspiration and purchase intent. Anyone searching for a particular cut, care instructions or a matching combination gets an answer in AI Overviews that often touches several brands and retailers at once.
That makes evaluation difficult. A drop in clicks can mean that the brand is losing visibility. But it can also mean that it is named in a summary and people later turn up directly in the online shop or in a store. These paths are hard to trace, and simple metrics do not capture them.
I already dealt with this question as Creative Director Digital at Breuninger. Between 2017 and 2020, analytics, attribution and SEO were part of my responsibility, together with a team of 10 people. Even then it was clear that a click only shows part of the effect. People who inform themselves online often buy in another place and at another time. Whoever measures visibility in search only by visits therefore underestimates its contribution. With AI Overviews this gap becomes bigger, but it is not new.
Then there is the quality of product information. AI systems draw on what they find. If details on materials, fit descriptions or care instructions are incomplete or contradictory, the answer becomes inaccurate. Clean product data management is therefore also a question of visibility.
Finally, the visual level plays a big role in fashion. A text summary cannot show how a fabric falls. That is a limit of the current formats and at the same time an opportunity for brands that work with good images and precise language.
What this means for a brand
For a brand this first means rearranging its expectations of search. Visibility in AI Overviews is a form of brand presence, even if it does not lead to visits straight away. It therefore belongs in the same consideration as other forms of awareness and not exclusively in the reporting of the performance channels.
Second, it pays to think of content from the question. Which questions do people ask before deciding on a product? Where can the brand answer with real expertise? Content that answers these questions clearly and honestly is equally useful for people and for AI systems. I have written more about this in the article on visibility in language models.
Third, you need control over what is shown. Google names the familiar means for this, such as nosnippet, data-nosnippet, max-snippet or noindex. This control should be used deliberately. Whoever blocks too much disappears from the answers. Whoever does not steer at all leaves the interpretation to others. The right balance depends on which content the brand regards as its actual capital.
And fourth, consistency pays off. A brand that uses the same terms on its website, in product data and in communication is more easily understood correctly.
Where I am still unsure
I am not sure how the relationship between search engine and website will develop in the long run. Today AI Overviews refer to sources. How much value these references have for brands is still hard to judge. The answer will probably differ greatly depending on the industry and the type of question.
I am also unsure about measurement. The tools for evaluation are developing, but they do not yet fully reflect the new reality. I think it is wise to work with several perspectives for a while: classic metrics, observing your own presence in the answers and signals from brand building. Whoever treats AI visibility as a topic of its own has an advantage.
On one point, however, I am sure. The foundations of good work remain valid. Clear content, correct data and a recognisable attitude are the best protection against any change in the search interface. In the end, AI Overviews reward what also convinces people: content that really answers a question.
For my own planning I draw a simple conclusion from this. I would not invest in special formats that follow a presumed logic of the machines. I would invest in the quality of my own content, in maintaining product data and in a language that is precise and unmistakable. That is less exciting than a new trick, but it lasts longer. And it works regardless of what search will look like in a few years.