AI Overviews and SEO: Debunking 6 Myths

SEO isn’t dead, and AI overviews aren’t traffic killers in themselves. The most sensational claims rarely stand up to fact-checking. Between panic and ignorance lies the sober reality. This article examines the six most common myths.

Ever since Google rolled out AI Overviews on a wider scale, there have been more opinions than hard facts circulating within the industry. On the one hand, there are those proclaiming the demise of traditional SEO; on the other, those who claim that nothing is changing anyway. Both positions can be tested against concrete data, and that is precisely what this article does, debunking each myth one by one.

Myth 1: SEO is dead

SEO is not dead; rather, its role is evolving because the same fundamentals that apply to traditional ranking also form the basis for visibility in AI-generated responses. This statement can be directly substantiated by an official source.

Google states in its own documentation AI Features and Your Website It is clear that there is no separate set of rules for inclusion in AI Overviews. The same basic principles apply as for regular ranking: technical quality, relevant content and authority. The ‘SEO is dead’ narrative confuses a shift in presentation with the end of the underlying discipline.

It’s easy to see where this myth comes from. Any visible change to the search results page instinctively triggers doomsday scenarios; this has been the case with previous Google updates too. In fact, AI Overviews do not alter the fundamentals of good visibility; they simply add an additional format on top. Anyone who wants to continue to be found in traditional results lists still needs to do exactly the work that has always defined SEO, supplemented only by the question of whether the same content is also suitable as an answer snippet.

A look at our own practice reveals the same pattern. Websites that already had a solid technical foundation, clear page structures and thematic depth before the introduction of AI Overviews generally feature in the new AI response formats as well. Websites that previously had weaknesses in terms of loading time, structure or clarity of content now lose out in two areas rather than just one. The difference, therefore, does not lie in a new discipline suddenly replacing everything, but in the fact that existing weaknesses are now doubly apparent.

This observation is consistent with a simple principle to bear in mind for any new search feature: every new format builds on existing indexing and ranking; it does not replace them. A Google system must first be able to crawl, understand and categorise a page before it can cite it in any form or display it in a answer box. It is precisely this sequence that remains unchanged, no matter how many new display formats are added to the results page.

Myth 2: AI overviews always cost traffic

AI overviews do not reduce the number of clicks across the board; rather, the effect depends heavily on the type of search query. For purely informational queries with a short, clear answer, the reduction is more pronounced than for more complex queries or those related to purchasing.

A study by the Pew Research Centres shows that users are less likely to click on links when an AI summary appears in the search results, although there are significant differences depending on the type of query. At the same time, the findings show Historical data from Semrush, that the visibility of AI Overviews has declined slightly in certain areas following an initial sharp rise, suggesting a levelling off rather than a sustained extreme trend.

For businesses, this means that for highly informational topics, it is worth taking a closer look at their own traffic trends; for transactional or highly brand-related search queries, the impact is generally much smaller. So anyone who makes a blanket claim that AI overviews are eating away at all traffic equally is ignoring precisely this difference between query types. A blog article addressing a general knowledge-based query is significantly more affected than a product page that someone is searching for with a clear intention to buy.

In practical terms, this means that, when it comes to your own content strategy, you should segment traffic by query type rather than looking at a single overall figure. If you find that only highly informational advice pages are showing declines, whilst product and service pages remain stable, you can respond in a targeted manner – for example, by linking informational content more closely to a clear next step, rather than focusing solely on the acquisition of knowledge.

The trend over time also warrants closer examination. The Semrush data does not show a linear, ever-accelerating decline, but rather a wave-like pattern: first a sharp rise in AI Overview visibility, followed by a period of consolidation. So, anyone reading the latest figures on traffic losses should always bear in mind when the data was collected; a single data point says little about the long-term trend.

Myth 3: AI systems only cite big brands

AI systems do not primarily cite sources based on brand size, but rather on content type and clarity, which means that even smaller providers regularly appear as sources. What matters is the way the content is presented, not the advertising budget behind it.

One Analysis of more than 75,000 AI responses shows that comparative list articles account for around 32.5 per cent of the top citations, which is significantly more than brand awareness alone would explain. The key factors in this study were, above all, whether the content was clearly structured, presented in a comparative format and backed up by clear evidence. A specialist website featuring a precise, well-structured comparative article therefore has a real chance of being cited in an AI response, regardless of whether it is run by a global corporation or a specialised SME.

For small and medium-sized brands, this is one of the most encouraging insights regarding GEO. Whilst traditional ranking for highly competitive keywords often depends heavily on domain authority and a long history, AI-generated snippets place greater emphasis on the immediate quality and clarity of individual content. This shifts the focus back somewhat towards content creation, away from pure reach, and gives specialist websites with limited budgets but high-quality, precise content a genuine chance of visibility in generative answers.

This does not mean that brand awareness plays no role at all; it simply works differently to how it does in traditional rankings. A well-known brand is more often cited as additional validation when content is otherwise of equal quality, but it does not replace the fundamental requirement for clearly structured, well-sourced content. Anyone who believes that their own brand awareness alone is sufficient for GEO, and who neglects the structure of their content, will be overtaken by smaller but more professionally presented competitors.

Myth 4: GEO is just a buzzword

GEO is not merely a marketing buzzword, but a term with clear academic origins, which refers to a study published in 2023 Research conducted by Princeton University is on the decline. It is this origin that sets GEO apart from many other short-lived industry terms.

The underlying study systematically investigated which content-related adjustments increase the likelihood of being cited as a source in generative AI responses, and demonstrated measurable effects of such adjustments. The fact that an entire landscape of consultancy services and tools has since emerged does not detract from the solid foundation of the concept itself. Anyone who dismisses GEO as mere hype fails to recognise that the underlying shift in search behaviour is real and measurable, regardless of what it ends up being called.

The accusation that it is merely a buzzword usually arises when agencies use the term to describe everything under the sun, from pure keyword optimisation to vague promises about AI visibility. This dilution is a real problem within the industry, but it says nothing about the scientific basis of the concept. Anyone seriously engaged in GEO should therefore make a clear distinction between its solid academic core – where clear structures and up-to-date, evidence-based content increase the likelihood of being cited – and the marketing jargon that has since grown up around it.

In practice, a simple test helps to distinguish reputable GEO offerings from mere buzzword marketing: Can a provider explain in concrete terms what structural or content-related adjustments they are making and why these are intended to increase the likelihood of being cited, or do they merely make vague promises about general AI visibility? The Princeton study itself identifies specific, verifiable factors; it is precisely this level of specificity that allows one to distinguish reputable work from mere trend-driven marketing.

Myths 5 and 6: AI traffic is worthless; everything stays the same

AI traffic is not worthless; in fact, according to the available data, it tends to convert even better than traditional search traffic, and the claim that search behaviour remains fundamentally unchanged does not hold up in the light of the current data either.

Various Analyses of the conversion performance of AI-generated traffic achieve a conversion rate that is roughly four to five times higher than that of traditional organic traffic. One possible explanation is that users who arrive at a website via an AI-generated response have already largely completed their initial research and arrive with a clearer purpose, rather than still comparing several results side by side. At the same time, the sharp rise in usage figures for AI search tools shows that search behaviour is indeed changing structurally, not just superficially.

This higher conversion rate has practical implications for the evaluation of marketing budgets. Anyone who simply ignores AI referral traffic in their own performance metrics or classifies it under ‘Other sources’ may be systematically underestimating the value of certain content investments. An article that generates little traditional search traffic but an above-average number of AI citations can still make an above-average contribution to business results if it is measured correctly.

Both extremes – “AI traffic is useless” and “nothing will change anyway” – can thus be equally refuted. Anyone who ignores AI traffic because they consider it inferior is missing out on a target audience that is more likely than average to convert. At the same time, anyone who believes their own content marketing can carry on unchanged as it did two years ago is underestimating just how much the ways in which people actually find their way to a website today have changed.

For your own business, this means it’s worth tracking AI referral traffic separately, rather than letting it disappear unnoticed within general direct traffic or organic traffic. Only by measuring it can you assess the actual contribution this new traffic source makes to your business’s success. You can read more about how to implement this in how GEO really works and in a direct comparison of SEO and GEO. Anyone wishing to explore the subject in greater depth will find in the full GEO Guide the detailed strategy behind it.

If we summarise these six myths, a clear picture emerges: neither panic nor ignorance are appropriate responses to current developments in search. SEO remains the foundation; AI overviews do not affect every search query equally; citations depend on content quality rather than brand size; GEO has a solid foundation; and AI traffic tends to be even more valuable than its reputation suggests. Those who are aware of these facts can allocate resources where they really make a difference, rather than reacting to headlines.