The top of a Google results page stopped being a list of blue links. In its place comes a finished block of text, written by artificial intelligence, with a few discreet credit links off to the side.
The shift is no longer an experiment. According to Similarweb, more than 43% of search sessions in the United States already show an AI overview, against roughly 15% a year earlier.
For anyone producing content, the AI-Powered Overview reorganizes the whole contest. The question stopped being which position you hold and became whether your text was read, summarized and credited inside the answer.
The cost of being left out has a known size. A Pew Research Center study found that with an AI summary on screen, clicks on a traditional organic result drop from 15% to 8% of visits.
What changes in practice, then, is not the SEO technique. It is the standard of clarity, structure and sourcing demanded of every passage you publish.
What is an AI-Powered Overview?
To find out in practice what an AI overview is, type into Google, for example, "how to write better" and notice that a collection of answers on this subject will be automatically generated for you.
These answers are generated based on a collection of information from different sources available on the web. Note that next to the answers, there are also links to access the sources of the research and if you think it necessary, analyze the full content.
With the AI overview, Google aims to improve the search experience by providing users with quick answers in one place, without the need to search multiple sites.
The feature started as an experiment. It was tested inside Search Labs under the name Search Generative Experience, or SGE, before Google rolled it out to the general public in the United States in May 2024, already renamed AI Overviews.
The name matters when you go looking for information. Content that still says SGE is describing a closed phase of the product, and Google documentation today files the subject under AI features in Google Search.
Caption: The summary now occupies the space the first position used to own, and the contest became about being cited.
How does Google decide when to show an overview?
It's clear from Google's commitment that AI overviews can redefine the way information is delivered to its users and will also change search in a similar way to what happened with featured snippets, which is that highlighted box with the questions asked by users.
One of the most important points about the AI overview is that its use can lead to a more complete search, with many fewer clicks. This is because users will find information without having to visit other sites.
The effect on organic traffic has left the realm of prediction. A Pew Research Center study published in July 2025 measured that users who met an AI summary clicked a traditional result in 8% of visits, against 15% when no summary appeared.
Clicking inside the overview itself is rarer still. In the same study, only 1% of visits ended with a click on one of the source links displayed beside the summary.
On the other hand, if the content is featured in a Google AI summary, this could increase the visibility of your site and your click-through rate (CTR).
Another benefit of the AI-based overview is that it generates more qualified traffic to websites. Simply because users, by clicking on the sites listed as sources in an AI overview, can become more interested and involved in the content to which they have already been introduced.
What is the difference between an AI overview and AI Mode?
An AI-Powered Overview is a block inside the traditional results page. AI Mode is a separate surface, announced by Google in May 2025, where the whole exchange happens with the AI and the list of links stops being the centre of the page.
Both run on the same base. Gemini writes both answers from passages retrieved by the core ranking systems of Search, which means there is no separate optimization for one and for the other, only Generative Engine Optimization applied to both.
The difference that matters for content teams is depth. AI Mode takes longer, chained questions, which raises the number of subquestions generated and favours pages with broad coverage of the topic.
In practice this moves the planning target. Instead of one post per keyword, what sustains presence across both surfaces is a set of pieces that covers the subject from several angles.
How does the AI overview work?
AI overviews appear in search results when Google identifies that answers generated by Artificial Intelligence may be useful to users. However, they do not appear in all searches.
Google displays this information selectively, prioritizing more complex queries that would normally require several searches to answer.
Complex queries tend to become repeated queries. The SemRush zero-click study recorded that 17.9% of desktop searches and 29.3% of mobile searches were users typing a new keyword right afterwards, a sign the answer did not resolve.
As they rely on AI, there is a possibility that the answers generated are wrong or biased. Google has already stated that it is working to minimize these problems.
Advertising inside the overview is also on Google's radar, marked as sponsored content. Check the official Search documentation before planning media around that space, since the format keeps evolving.
What do you actually control in an AI-Powered Overview?
Three of the most repeated questions about AI overviews changed answers between 2024 and 2026. Worth checking each one before deciding anything about presence in that space.
| What the post claimed | Where it stands today | Where to confirm |
|---|---|---|
| You cannot keep content out of AI answers | You can | Search generative AI control, in Search Console |
| Google separates no AI data at all | It shows impressions | Generative AI performance report, since June 2026 |
| The feature is called SGE | It is called AI Overviews | Google I/O announcement, May 2024 |
| It appears in 0.55% of US searches | More than 43% | Similarweb, 2026 data |
Table: What changed in the AI-powered overview between the launch of the feature and today.
Can you keep your site out of AI overviews?
Yes, since June 2026. The Search generative AI control in Search Console settings offers include, exclude or inherit. Google states the choice is not used as a ranking signal in traditional Search.
The cost is explicit in the documentation itself: sites that opt out receive no traffic and no impressions from generative features. For most operations, opting out closes a door without opening another.
Does blocking Google-Extended solve it?
It does not, because it governs something else. Google documentation is literal in saying that Google-Extended does not impact a site's inclusion in Google Search and is not used as a ranking signal there.
What it governs is training and grounding for Gemini and Vertex AI. To limit what shows from your pages in Search, the controls are different ones: nosnippet, data-nosnippet, max-snippet and noindex, the same ones that apply to People also ask.
Does Search Console report overview performance?
It reports impressions, not clicks. Since June 2026 there is a generative AI performance report broken down by page, country, device and date. There is no dimension that separates AI Overviews from AI Mode.
In practice you can see whether presence is growing or shrinking, not how much traffic it sends. Until Google opens up clicks, reading the result stays indirect.
Did overviews replace featured snippets?
They did not. The official featured snippets documentation is still live and was updated in December 2025, with its own control guidance, which points to a feature that is alive and coexisting with the overview.
How do you optimize content for an AI-Powered Overview?
Although there are no specific optimization techniques for AI overviews, Google has stated that its systems automatically choose the links and sources that appear in these overviews.
Therefore, the key is to follow Google's general SEO guidelines, such as those described in Google Search Essentials.
If you already apply good SEO practices, you're on the right track. Below, we highlight the main practices for optimizing your site:
Tip 1: Make sure your pages are crawlable and indexable
To enter an AI-Powered Overview, a page has to be crawled and indexed first. The first thing to check is the robots.txt file: confirm it is not blocking access to the pages you want to see cited.
Check that there are no meta tags that accidentally prevent indexing. Take the opportunity to correct pages with 404 or 403 errors to avoid crawling problems. Finally, organize your pages clearly to make it easier for Google's crawlers.
Tip 2: Use long-tail keywords
Overviews created by AI are most useful for complex queries, often based on long-tail keywords, i.e. those with more than four words. Once you've identified specific terms, create optimized content that matches these searches.
Tip 3: Create high-quality content
Quality content has a greater chance of appearing in Google overviews created by AI. But what does quality content mean? It means that:
- They need to be accurate, with reliable and verified information;
- They must be original, with unique ideas or an innovative angle on the information;
- They should be produced by experts in the field; and
- They must offer value to the user, meeting the real needs and questions of the audience.
Outside sensitive topics, what moves the needle on AI citation is something more specific: named sources, verifiable statistics and expert quotations. Those three tactics are the ones with measured effect in the academic literature on generative search.
Tip 4: Improve the user experience (UX)
A good user experience can also influence your visibility. Pay attention to speed and ensure that your pages load quickly. Use responsive design to offer a good experience on any device and protect user data with HTTPS.
Tip 5: Apply on-page SEO techniques
Optimize the technical side of on page SEO so Google can read your content more easily. These are cheap adjustments that pay off in traditional search and in AI answers at the same time.
- Title tags: include relevant keywords in titles;
- Meta descriptions: create attractive and informative descriptions;
- URLs: use short, descriptive URLs;
- Headings: structure your content with clear headings(H1, H2, H3);
- Keywords: use keywords strategically, but avoid overdoing it.
Tip 6: Build backlinks
Backlinks are links from other sites pointing at yours, and they work as an authority signal. A domain with a solid record of external citation tends to be picked more often as the source of an AI overview.
Reach out to sites in your sector with the value of your content in hand. That outreach is the core of any off page tactic worth the time it takes.
By applying these practices, you increase the chances of your site being considered in AI overviews and improve your presence in search results.
The length of your content also makes a difference in rankings! Find out the ideal length for a blog post to stand out on Google and achieve better positions in search results.
Frequently asked questions about AI-Powered Overviews
So where do you start to become a cited source?
Start with the pages that already get search volume and are now losing clicks to the summary. Those are the URLs where the performance report shows high impressions sitting next to a low click count.
On each one, do three things in order: open every section with the direct answer, swap generic claims for data with a named source, and split paragraphs that stack more than one idea, without fixating on the ideal post length.
After that, watch the generative AI report in Search Console. It shows impressions across AI surfaces and is currently the only official reading available on this kind of presence.
If your content marketing operation depends on organic search and you want to understand which pages still have a shot at becoming a cited source, talk to the mkt4edu team.
Portugal