Google AI Overviews: Why Being First Is No Longer Enough in 2026

For years, the rule of organic search was straightforward.

A business aimed to reach Google’s first page. Then it targeted the top three. The first position was the ideal outcome. The higher a page ranked, the more visibility and clicks it received.

This logic still holds. But it’s no longer enough.

Since July 22, 2026, Google’s AI Overviews have been officially available in France. Google can now display an AI-generated answer before traditional organic results. This answer summarizes multiple sources and may satisfy the user’s query without them visiting a single site.

A page can thus remain first on Google and still lose some of its traffic.

It’s still well-ranked. Its content remains relevant. No competitor has overtaken it. Yet, Google’s generated block takes up significant screen space and answers the question directly.

The shift is profound.

The challenge is no longer just about ranking well. It’s also about being understood by Google’s systems, selected as a source, cited in the answer, and recognized as a subject-matter authority.

Google’s AI SEO thus adds a new layer to traditional search optimization. This layer is often called GEO, for Generative Engine Optimization.

How AI Overviews are changing Google search

An AI Overview is an answer produced by Google’s artificial intelligence models.

When Google determines that this answer provides added value, it may display a summary at the top of the page. This summary includes explanations, advice, or comparisons, along with links to certain sources.

Users no longer see just a list of results. They already receive an answer.

Consider a search like:

"Which invoicing software should a five-employee small business choose?"

In a traditional results page, the user would open multiple articles. They’d compare features, prices, and reviews, then return to Google to check another result.

With an AI Overview, Google can directly present the criteria to check. It can compare different types of solutions and explain key considerations. The user gets an initial summary without leaving the page.

Google states that AI Overviews appear when its systems determine that a generated answer provides greater value than traditional results. The company also notes that links in these answers should allow users to explore the topic further.

This stance is public and worth reading in full, as it contradicts much of what’s being sold today: Google explicitly states that "there are no additional requirements to appear in AI Overviews or AI Mode, nor any special optimization needed," and that "SEO best practices remain relevant" for these features. In other words, no one can sell you a secret technical lever to enter generated answers. Source: Google’s guide on AI features and your site.

The problem for businesses is clear.

The AI Overview captures attention before organic results. Even when a page remains first, users may feel they already have enough information.

They no longer click.

The drop in clicks is no longer just a hypothesis

Multiple studies observe a significant shift in user behavior when Google displays an AI-generated answer.

The Pew Research Center analyzed 68,879 Google searches conducted in the US in March 2025. When an AI summary appeared, users clicked on a traditional result in just 8% of visits. Without an AI summary, this rate reached 15%.

Methodological details matter when interpreting these figures: the study tracked the browsing behavior of 900 US adult volunteers, 58% of whom performed at least one search that triggered an AI summary in March 2025; of the 68,879 searches in the dataset, 12,593 generated a summary, and result pages were collected from April 7 to 17, 2025. The same study also measured a third, less-discussed but decisive effect: users ended their browsing session in 26% of cases after viewing a page with an AI summary, compared to 16% without. Full study: Pew Research Center analysis of Google’s AI summaries.

Links placed within the summary itself received very few clicks. According to the study, an AI Overview link was opened in just 1% of relevant visits. Users also ended their sessions more often after reading the generated answer.

A 2026 field experiment reached a similar conclusion. When AI Overviews appeared, outbound organic clicks dropped by approximately 39.8%. Zero-click searches increased by 34.5%. Researchers observed no equivalent improvement in user-reported satisfaction.

Ahrefs examined the impact on already ranked pages. Based on December 2025 data, the presence of an AI Overview reduced the click-through rate of the top-ranked result by roughly 58%. The estimated impact then decreased with ranking, though it remained negative up to the tenth position.

Another study leveraged the gradual rollout of AI Overviews to assess their impact on Wikipedia. Researchers estimate that exposure to AI answers reduced daily traffic to analyzed English-language pages by about 15%. The effect varied by topic. Content that was easy to summarize appeared more affected than subjects requiring in-depth exploration.

These three studies deserve precise identification because their methodologies differ, and it is this difference that makes their convergence so compelling. The field experiment was conducted by Saharsh Agarwal and Ananya Sen: 1,065 Chrome users were randomly assigned to perform Google searches with or without AI Overviews via a dedicated extension, the first truly randomized protocol on the subject. It measured a −39.8% drop in outbound organic clicks and a +34.5% rise in zero-click searches, with no effect on sponsored clicks or search frequency (study available on SSRN). Ahrefs’ figure comes from its 2026 study update, covering 300,000 keywords and December 2025 data: the CTR of the top position fell by about 58%, compared to 34.5% in the initial March 2025 measurement, the gap between the two editions is itself the most telling insight (Ahrefs study update). Finally, the Wikipedia study exploited the staggered geographic rollout of AI Overviews, comparing 161,382 pairs of matched articles: English versions exposed to AI Overviews versus their Hindi, Indonesian, Japanese, and Portuguese equivalents, which were not yet exposed. This approach establishes causality rather than mere correlation (research paper on arXiv). One nuance from this last study is worth noting for any publisher: the decline was most pronounced for cultural content and far less significant for scientific and technical topics. The more your subject can be summarized in three sentences, the more exposed you are.

These figures should not be applied uniformly to every SME.

The outcome depends on the industry, the query, the type of content, and the Google results page layout. An urgent local search does not produce the same behavior as a general question. A quote request is not the same as a definition search.

The methodologies used in the studies also differ.

But their overall direction is consistent: when Google provides a direct answer, the likelihood of a click decreases.

For a business leader, this means a drop in traffic does not necessarily indicate a drop in ranking.

A page may retain its position while receiving fewer visits.

Why the top position no longer guarantees a citation

Here lies the most critical point.

The sources selected by Google’s AI Overviews are not always the pages ranked highest for the user’s query.

Google states that its AI features rely on its search index and ranking systems. Traditional SEO thus remains a critical foundation. A page must be indexed and eligible to appear with a snippet to be used as a link in Google’s generative features.

Yet Google is not necessarily limited to the ten visible results for the initial search.

The engine uses a method called “query fan-out.” The model breaks down the query into multiple sub-questions, then runs several related searches to build its answer.

A query such as:

"How to choose a CRM suited to an industrial SME?"

can trigger internal searches on:

• user count;
• ERP integration;
• sales tracking;
• quote management;
• deployment cost;
• data security;
• team training.

Google may then select a highly specific source on CRM and ERP integration, even if that source does not appear in the top 10 for the main query.

An Ahrefs study published in March 2026 analyzed 863,000 result pages and roughly four million cited URLs. Only 37.9% of URLs in AI Overviews appeared in the top 10 blocks for the corresponding initial query.

Around 31% ranked between positions 11 and 100. A nearly identical share did not appear in the top 100 blocks for that query at all.

An academic study of 55,393 searches in 2026 also found that nearly 30% of domains cited in AI Overviews did not appear in the first-page results displayed alongside them. The authors concluded that the source selection mechanism does not fully align with the visible organic ranking.

There are thus two related but distinct competitions.

The first is to rank well in organic results.

The second is to become a useful source for the generated answer.

A page can win the first competition and lose the second.

AI Overview SEO does not replace traditional SEO

Faced with this shift, some businesses might be tempted to abandon traditional SEO in favor of GEO.

This would be a misreading of the situation.

Google clearly states that its generative features rely on its search systems and index. SEO fundamentals thus remain essential: crawling, indexing, content quality, internal linking, technical performance, and reading experience.

Without a solid SEO foundation, ranking in AI Overviews becomes far more difficult.

A poorly indexed page won’t be a strong candidate. Content buried in a complex interface will be less accessible. A slow or hard-to-read page will struggle to deliver value, even if it contains the right keywords.

GEO does not replace SEO.

It expands it.

Traditional SEO primarily aims to make a page visible and competitive in search results.

GEO also seeks to make information easy to select, verify, and integrate into a generated response.

This evolution can be summarized in four steps:

1. Be accessible to Google.
2. Be relevant to the topic.
3. Be selected as a source.
4. Be distinctive enough for the brand to be recognized.

The first two steps belong to the SEO foundation. The latter two carry more weight with Google AI SEO.

Traffic is no longer the sole visibility metric.

For a long time, SEO teams tracked a few simple indicators:

• average position;
• number of impressions;
• click-through rate;
• organic traffic;
• conversions from Google.

This data remains useful. But it no longer captures the full scope of Google AI visibility.

Imagine a page stays in second position for a strategic query. Its impressions remain stable. Yet, its clicks drop by 25%.

A traditional analysis might suggest several hypotheses: a less compelling title, a new competitor, a weaker season, or a shift in user intent.

The emergence of AI Overviews adds another explanation.

Google now answers before the page.

Conversely, a site may be cited in a generated response without receiving many clicks. It gains brand presence, but this presence doesn’t immediately appear in Google Analytics.

Leaders must therefore distinguish between three types of visibility.

Ranking visibility

The site appears in organic results and holds a measurable position.

Citation visibility

A page from the site is used as a source in an AI Overview.

Brand visibility

The company’s name, product, method, or expertise becomes identifiable in the response or in the suggested sources.

The third type is the hardest to measure. Yet it remains important.

When AI answers directly, users may remember a brand before even visiting its site. They may also search for it later, find it on LinkedIn, or discuss it with a colleague.

The absence of an immediate click does not always mean the absence of value. However, a mere appearance among several links does not guarantee genuine memorability either.

The goal is to become an identifiable source, not just a discreet URL.

What Google truly expects from content in 2026

Google’s official stance remains fairly restrained.

To appear in generative features, there is no secret tag or mandatory format. Google recommends continuing strong SEO practices and producing useful, reliable content designed for people.

The company emphasizes content that is hard to replace.

A page that repeats the same advice as twenty competitors offers little unique value. It may be correct, well-structured, and well-optimized. Yet Google already has many sources saying the same thing.

Conversely, content becomes more distinctive when it includes:

• real-world experience;
• a precise method;
• company-specific data;
• a comparison based on testing;
• a detailed client case study;
• a clearly attributed expert opinion;
• recent, verifiable information.

Google refers to “non-commodity” content, content that cannot be replaced by a generic summary. Its documentation advises providing an original perspective and avoiding recycled material already available on the web.

For an SME, this is good news.

A small business may not always publish as much as a large group. However, it can produce more precise, concrete, and field-relevant information.

An accounting firm can explain the real obstacles encountered during a software transition.

An industrial company can publish a calculation method derived from its projects.

An agency can document the choices made during a real mission.

A craftsman can highlight the signs that help identify a problem before an intervention.

This type of content fuels both traditional SEO and AI Overview ranking.

How to adapt a content strategy to Google AI SEO

A serious GEO strategy does not start with writing dozens of new articles.

It begins with an analysis of the topics on which the company must be recognized.

Start with the decisions made by the customer

A Google search does not always represent a simple request for information. It may conceal a decision.

The user might be looking to choose a provider, compare two solutions, set a budget, or mitigate a risk.

An effective content strategy must therefore address the full decision-making process.

Let’s take an SME that sells field service management software.

The keyword “technician intervention software” may seem central. But the AI Overview can explore multiple aspects:

• scheduling;
• movement tracking;
• work order signing;
• billing;
• offline mode;
• inventory management;
• price per user;
• data migration.

A general page may not always cover these subtopics with sufficient depth.

SEO and GEO work involves building a coherent set. A main page covers the broad topic. Supporting content provides more precise answers. Case studies prove real-world use.

This structure expands topic coverage without creating artificial pages for every keyword variation.

Produce information Google cannot find everywhere

A general article on 'the benefits of a CRM' will be easy to summarize. It will also be easy to replace.

An analysis of three CRM deployments in SMEs with fewer than fifty employees offers more value.

Content can include:

• average project duration;
• the longest tasks;
• framing errors observed;
• features rarely used after six months;
• the criteria that truly influenced adoption.

These details enhance the content’s usefulness. They also give Google precise facts that can support an answer.

This is not about inventing figures to appear expert.

Data must be accurate, explained, and contextualized. A sample of five clients can be useful if its size is clearly stated.

Make answers easy to identify

Google states that it is not necessary to split every text into tiny fragments for AI systems. There is also no imposed ideal length. The priority remains clarity for the reader.

This does not mean a page can be confusing.

A key question must receive a visible answer. Headings must reflect the actual content of the section. Technical terms must be explained. Comparisons must rely on stable criteria.

A good page can start a section with a direct answer, then provide details and nuances.

This structure helps the rushed reader. It also makes it easier to identify the relevant passage.

Strengthen proof around the author and the company

Information inspires more trust when it is clear who is publishing it.

The page should therefore specify the author, their role, and their experience. Cited studies must be accessible. Update dates must be consistent. Examples must be attributed to a source or presented as anonymized cases.

Proof is not limited to the page.

A brand recognized across multiple channels has a richer context. It may be mentioned in industry media, a professional association, an interview, a podcast, or a video.

Google states that its AI features can use content from pages, videos, forums, and other web spaces. However, it advises against seeking false mentions or artificially manipulating this presence.

The right goal is not to multiply valueless citations.

The focus should be on building a consistent reputation.

Leverage multiple formats when the topic justifies it

AI Overview SEO does not rely solely on blog posts.

Google can also display images, videos, local business information, and product data. Its documentation recommends continuing to use Google Business Profile and Merchant Center when these tools align with the business activity.

Ahrefs also observed a strong presence of YouTube in the sources used by AI Overviews. In its March 2026 study, YouTube accounted for 18.2% of citations that did not appear in the top 100 results of the direct query.

An SME does not need to become a video media outlet.

But it can repurpose a strategic piece of content into multiple resources:

• a comprehensive article;
• a video demonstration;
• an FAQ;
• a client case study;
• a service page;
• a diagram or comparison table.

Each format addresses a different need. Together, they deepen subject understanding and enhance Google AI visibility.

The central role of the brand in GEO

Traditional SEO has sometimes led companies to publish very neutral content.

The brand name was subtle. The article primarily aimed to match the keyword. The logic was simple: secure the click, then introduce the company.

AI Overviews make this approach more vulnerable.

When Google absorbs part of the answer, the click can no longer be taken for granted. The brand must therefore be embedded in the value delivered.

This does not mean repeating its name in every paragraph.

Instead, the brand should be linked to something specific:

• a methodology;
• a study;
• a tool;
• an analytical framework;
• sector-specific expertise;
• regularly updated data.

A recruitment agency can publish a local salary barometer.

A consulting firm can create an evaluation grid.

A construction company can document a diagnostic method.

An IT company can publish an annual index of digital maturity for SMEs.

This type of editorial asset can be cited, reused, and remembered. It gives the company an identity in generated responses.

GEO is therefore not just about optimising pages.

It also involves making expertise attributable.

Measuring AI Overview SEO without misidentifying KPIs

A Google AI SEO strategy must be tracked with multiple indicator families.

In June 2026, Google announced new Search Console reports dedicated to generative features. These reports provide access to impressions, visible pages, countries, devices, and trends over time. Their rollout was initially limited to a subset of sites.

When available, these reports provide a useful baseline. They do not replace manual analysis.

A company must also monitor the responses generated for its priority queries.

Check:

• whether an AI Overview appears;
• which sources are selected;
• which angle is used;
• whether the brand is cited;
• which competitors are visible;
• what information seems missing from the site.

This monitoring should cover query groups, not just exact keywords.

The query fan-out requires thinking in terms of topics and sub-questions.

GEO shortcuts that waste time

As with any major change, quick-fix solutions soon emerge.

Some claim to make a site appear in AI Overviews with a specific file, a special tag, or a repeated structure across all pages.

Google has stated it does not use the llms.txt file to improve a site’s presence in its generative features. The company also clarifies that no special structured data is required to appear in an AI Overview.

Structured data remains useful for Google’s classic functions. It must match the visible content. But it is not a GEO shortcut.

Google also advises against creating a large number of pages to artificially cover every variation of a query. This approach can produce low-quality content and fall under its rules against large-scale page generation.

Ranking in AI Overviews cannot be achieved by simply adding another layer of keywords to generic texts.

It requires a deeper understanding of the customer, genuine expertise, and a coherent editorial architecture.

A realistic action plan for an SME

An SME does not need to overhaul its entire site in a few weeks.

It can start with topics that directly influence how its offering is understood.

Step one: identify exposed queries

Analyse the keywords already driving traffic and check which ones trigger an AI Overview.

Long-tail queries, phrased as questions or comparisons, are often strong candidates. The Pew Research Center’s study found that longer searches and full sentences more frequently triggered an AI summary in its March 2025 sample.

The goal is to identify pages likely to retain their rankings while losing clicks.

Step two: compare selected sources

For each key topic, examine the sites cited by Google.

The comparison should not be limited to traditional rankings.

A selected source may have a more recent page, more precise data, a clear video, or better coverage of a subtopic.

This analysis often reveals editorial competitors invisible in standard SEO tracking.

Step three: strengthen priority content

Next, enhance pages that already have a solid foundation.

The work may focus on:

• answer accuracy;
• original data;
• examples;
• the author’s expertise;
• sources;
• complementary formats;
• links to other pages on the site.

The update must provide genuine value. Changing a date or adding generic lines is not enough.

Step four: build external proof

Google AI visibility also depends on what the web says about the company.

Partnerships, appearances in trade media, detailed testimonials, and expert contributions can strengthen this presence.

The goal remains consistency.

An SME should be associated with a few specific topics rather than trying to appear everywhere.

Step five: track traffic and brand presence

A drop in clicks should be compared with changes in rankings, impressions, and presence in AI responses.

You should also monitor searches that include the brand name.

An increase in branded searches may indicate that some users discover the company before returning later. This interpretation should remain cautious, as multiple marketing actions can produce the same effect.

2026 SEO aims for presence, not just position

Google AI Overviews do not make organic search obsolete.

They change what it means to be visible.

A top position remains a valuable asset. It increases the likelihood of discovery and supports the site’s credibility. But it no longer guarantees the same click volume. Nor does it ensure selection as a source in the generated response.

AI Overview SEO therefore requires businesses to broaden their strategy.

You must continue to optimise the site. You must also create content precise enough to be reused, reliable enough to be cited, and distinctive enough to be associated with a brand.

The shift from SEO to Google AI SEO can be summarised as follows:

Previously, companies mainly sought to secure a ranking.

Today, they must secure a ranking, a citation, and recognition.

The top spot remains useful, but it is no longer the true peak

For SME and VSE leaders, the main risk is focusing solely on Google rankings.

A tracking dashboard can stay green while clicks decline.

The site hasn’t necessarily lost its ranking. The engine has simply changed how it distributes attention.

The right approach isn’t to replace the entire SEO strategy with a set of GEO tactics. It’s about checking where the company is visible, which sources Google retains, and which content can still build brand preference.

An SEO/GEO audit connects these three levels: organic ranking, presence in AI Overviews, and the brand’s editorial strength.

Agence Scroll helps businesses navigate this analysis. The goal is to identify exposed content, topics where the brand can become a reference, and editorial actions with lasting impact.