The rise of AI chat has shaken up the search engine market. It’s possible that the introduction of AIO is the first step in revolutionizing the way we search.
What is AI Overviews (AIO) and how will it affect SEO and search results?
In March 2025, Google made AI Overviews available in Poland in its search engine (translated into Polish as Przegląd od AI), i.e. answers to user questions generated by AI, presented above classic search results. Find out how AIOs look and work, and whether they herald a revolution in the use of Internet search engines and (yet another) death of SEO.
How AIO Score Works
AI overviews are search results generated using Google’s AI Gemini model. The goal of AIO is to present a synthesized knowledge pill for a given query, without having to familiarize yourself with the individual contents of external services (these, however, are the source of knowledge for the result generated in this way). In theory, Google is no longer limited to just pointing us to sources from which we can obtain some knowledge (if we familiarize ourselves with them), but carries out the entire process of data analysis and formulating answers (it does a kind of query/research for us and formulates an exact answer to the given topic based on it).
Initially, in Poland Przegląd od AI appeared only to logged-in and adult users (which could have somewhat limited its impact on the websites’ results), but now AIOs appear quite commonly, also to people who are not logged in and in the incognito mode of web browsers.
What does AIO look like?
At this point, most Google search users have probably encountered AI Review while searching for information on the web.
AI Overviews are displayed at the very top of search results above classic organic results, and also, interestingly, above paid results (this will probably change soon, because ads lose too much on this).
In the desktop version, a box with the AI answer appears under the query bar, on the left side we get the content of the answer with anchors to sources related to it, and on the right side we get links and teasers of source services, for the entire answer or for a specific fragment of content related to a given anchor. It is worth emphasizing here that the number of anchors to sources in the AIO result can be significant, and each anchor represents a slightly different set of sources. In such a system, additional mini-rankings of AIO sources appear inside the search engine, which complicate the classic understanding of a single, relatively stable ranking.
It is similar for mobile devices, except that the result is adjusted to the size of their screen and the way they are used – the generated response content is on the first screen, while the sources are displayed only after clicking on their icons, located in the upper right corner of the AI Review or on the appropriate anchor:
The AI Review result itself, as well as the sources associated with it, are often collapsed by default so as not to overwhelm all the search results – at this point, AIO can already take up virtually the entire first screen of the SERPs:
AI Overviews’ Impact on SEO Results and Beyond
The appearance of AIO significantly disrupts the current layout of search results. Ads and organic results are pushed to the next (if not further, assuming that someone develops such a response to a full format) screen.
The inevitable drop in CTR
The natural consequence of introducing AI Overviews is that users, receiving all the information on a plate already at the search results level, will not feel the need to click on any additional links and go to external services (even the source ones). In such a situation, we observe a massive decrease in the number of clicks. The first observations (based on our own data and industry voices) confirm that as a result of introducing AIO, CTR in Poland dropped by an average of 25%. Of course, a lot depends on the type of phrase, industry, commonness of knowledge on the subject, etc.
The greatest impact is obviously visible on all kinds of informational queries, for which AI can demonstrate its eloquence, and much smaller in strictly product or navigational ones:
AIO – Google’s Shot in the Knee or a Flight Ahead?
The introduction of AIO is associated not only with a decrease in organic clicks, but also with a decrease in ad traffic, which from Google’s perspective is certainly not something positive (fewer clicks = less money from advertisers). However, Google certainly treats it as a kind of investment that will only start to pay off in the future.
Google’s reports clearly indicated that the introduction of AI Overviews significantly increased the number of Google search queries by 10%, so from the perspective of Google maintaining its dominance in the search engine market, it was a good move:
User trust in generative results
The quality, credibility and trust of users in generative results and AI-generated answers is still not yet established. Google has already had several false starts in this regard, and surprisingly in very sensitive areas such as health (for which even classic organic results have additional requirements related to algorithm updates such as EAT , extended to EEAT, or YMYL), as a result of which at one point it severely restricted the display of such results in the US, where the first AIO implementations were carried out. Over time, the quality of these results has improved, and this process will probably continue to improve (for quite simple queries, the generated answers are already quite reliable and trustworthy).
There is still the issue of learning and developing such models, because new (fresh) information from source services is added to the data on which they were built (which may involve certain risks). False information placed on a large number of such services may cause AI responses generated on the basis of such sources to contain similar errors, and therefore to be untrue (and here the circle closes and we return to “doing SEO” and enchanting the Internet with sponsored articles).
AIO is not the same ranking as Google organic
Having the first organic position in classic search results does not necessarily mean that we will be included as a source of the AIO result (although of course it significantly increases our chances. It is estimated that by about 25%). Interestingly, AI Overviews often appear together with the current position 0 of search results, i.e. Direct Answers, and the DA source does not have to be included as an AIO source at the same time.
Google itself admits that AI Overviews and classic results are slightly different rankings, guided by slightly different rules and guiding principles.
The differences between AI Overviews and classic results were neatly summarized by ZipTie, emphasizing a significant change in the paradigm – a shift from presenting the best page (presenting the most information on a given topic) to presenting the best answer (precisely answering a given query and able to aggregate information from different subpages):
“Traditional SEO optimizes for the “best page.” When someone searches for “iPhone 15,” your page needs to cover everything to satisfy the user intent – specs, reviews, where to buy, comparisons.
AI Search looks for the “best answer.” When someone asks specifically about iPhone 15 battery life, you may rank top 1 in Google, but AI doesn’t care about it if you don’t provide a precise, relevant answer to that exact question.” ii
You can read more about how Google AI reads, understands and analyzes pages for generative results in Adam Piotrowski’s article on the Bluerank blog: How does Google read your page for AI Overview?
User Perspective – AIO as a Remedy for SEO Clutter on the Internet
Although – for me as an SEO specialist – solutions such as AIO may seem like a harbinger of the twilight of the industry, from the user’s perspective it is a great functionality, perfectly satisfying their intentions (we move the dilemmas related to quality and credibility to the sidelines here) and filtering out low-value, over-optimized content, for which we (as SEO specialists) are sometimes responsible (let the one who did not feel irritated trying to find answers to a simple question in Google “how long to boil a soft-boiled egg” or “is today a shopping Sunday” end up scrolling through low-quality, absurd tirades about the history of eggs and the exegesis of the days of the week be the first to cast a stone).
The introduction of AIO is a kind of reset from this perspective, forcing a return to a stronger focus on the user’s intention and needs, and not tricks and practices aimed at bending Google to display a specific page in the first position of search results. We can delude ourselves that Google is guided by the idea: Make Search Great Again – except that it cannot afford to drain so much capital, so at the moment we are probably operating in a narrow window, before the monetization of AIO (and AI chats in general), which gives us the feeling that we once again have clean, uncluttered and relevant search results. This moment will probably not last too long, however.
AI Overviews – how to analyze results?
With the introduction of AIO, Google had to slightly reprogram its search results page, which, along with an algorithm update, led to most popular visibility monitoring tools (Ahrefs, Senuto) stopping collecting data. This problem was quickly fixed, but another, much more serious one arose – as a result of the implementation of AI Review, CTRs began to drop, views increased, and average positions did not change much – the previous method of data analysis ceased to make sense.
AIO results were difficult to analyze, not only due to the new way of formulating queries (more complex, new phrases without any specific popularity), but also due to the issue of additional mini-rankings of sources (it is not entirely clear how to count their positions, clicks and views).
Currently, the topic is still developing. Ahrefs and SemRush have already started implementing AIO visibility reports based on phrases currently in their databases. GSC reports (unfortunately) several of the AIOs as regular clicks, although there are already rumors that AIOs will get a separate report in this tool. GA4 allows additional configuration that gives us insight into the amount of such traffic (e.g. tracking URLs with the fragment “#:~:text=”, but it may not be precise enough).
The Future of Search Results – Moving from SEO to GEO, or Generative Engine Optimization
Sparktoto claims that in 2024, almost 60% of searches ended with no clicks, with the rest mostly taken by the organic channel (an audible sigh of relief from SEO experts), followed by additional search engine functionalities (maps, videos, graphics, news, etc.), and only at the end we have paid traffic (the share of which seems slightly underestimated in this comparison):
However, there is a visible tendency for Google to increasingly take over the next stages of the user’s path, by not letting them leave its search engine for external sites. AI chats will probably try to do this even faster, making all stages of the Customer Journey (including placing an order and payment) take place exclusively from the chat level (just tell me what you want to buy, and I’ll take care of the rest – all you have to do is pick up the package from your door). There are already voices (including from Google itself) that the entire Internet, as we know it today, will transform into one big database, in which we will be mediated exclusively by AI agents).
In such a scenario, SEO, or rather GEO (as Generative Engine Optimization) will also have to transform and adapt, finding ways to optimize the business (as a service, or maybe just a database, a collection of information?) to organically exist and attract the attention of the AI agent. To meet the challenges of tomorrow, we already offer Generative Engine Optimization (GEO) in our services. Audits and strategies for AI Overviews .
AI Overviews Are Just the Beginning, or SEO Dies Once Again
The death/end of SEO is being heralded quite regularly. With virtually every new Google idea related to search results, someone decides that this is the end of SEO and that it’s time to move on. While previous changes were more about minor adjustments to new requirements and adjustments to operations, AI chats, even in such a limited form as AIO, are game-changers and require a fundamental rethinking of certain assumptions and practices.
Google has already announced that it will soon be more actively promoting the AI mode (AI Mode – which is essentially Gemini chat), placing it in the first tab from the left in its search engine.
In other words, Google wants to transfer users of traditional search results to a search engine that operates on a generative principle.
In addition, there is image search (Google Lens), supported by the increasingly popular augmented reality (AR) devices.
Of course, Google is not doing this for free, it has already announced the introduction of ads to AI Overviews and AI Mode. In such a situation, the biggest losers are publishers who feed their information to all kinds of chats and language models, gaining practically nothing in return (because the traffic stays in the search engine/chat and is directed to those who pay Google more by buying ads from them).
All this means that we need to look broader and go a little beyond our channel, focusing on the broadly understood building of brand and authority at many points of contact with users.
Authority will likely become increasingly important in the near future, enabling it to appear as a source of information for AI results, while a strong brand will ensure a flow of users directly to our business, regardless of the channel they come from.
Summary – Conclusions and Implications for SEO and Organic Channels
- AI Overviews (AIO), or AI Summary, has caused quite a stir in the SEO industry, fully satisfying the information needs of many users already at the search engine level, and thus significantly reducing CTRs.
- The new SEO reality is the need to move from doing SEO for phrases/visibility (creating the best possible website), to doing SEO for users and their intentions (creating the most precise answer possible) and building authority and brand.
- The future of SEO may be optimizing for generative results created by AI – Generative Engine Optimization (GEO).
Sources:
SparkToro. (2024). 2024 Zero-Click Search Study: For Every 1,000 US Google Searches, Only 374 Clicks Go to the Open Web; In the EU, It’s 360 . Retrieved from https://sparktoro.com/blog/2024-zero-click-search-study-for-every-1000-us-google-searches-only-374-clicks-go-to-the-open-web-in-the-eu-its-360/
Schwartz, B. (2024). Google Search Zero-Click Study 2024 . Search Engine Land. Retrieved from https://searchengineland.com/google-search-zero-click-study-2024-443869
Southern, M. (2024). Google Links To Itself: 43% Of AI Overviews Point Back To Google . Search Engine Journal. Retrieved from https://www.searchenginejournal.com/google-links-to-itself-43-of-ai-overviews-point-back-to-google/546574/
Southern, M. (2024). Does Google’s AI Overviews Violate Its Own Spam Policies? . Search Engine Journal. Retrieved from https://www.searchenginejournal.com/does-googles-ai-overviews-violate-its-own-spam-policies/547020/
Google. (n.d.). AI mode in Search . Google Blog. Retrieved from https://blog.google/products/search/ai-mode-search/
Schwartz, B. (2024). Google Search Market Share Drops 2024 . Search Engine Land. Retrieved from https://searchengineland.com/google-search-market-share-drops-2024-450497
Google. (2025). Google I/O ’25 Keynote . Retrieved from https://www.youtube.com/live/o8NiE3XMPrM?t=468s
KPP Marketing. (n.d.). How to Track Traffic From AIO, Featured Snippets & PAA Results in GA4 . Retrieved from https://kpplaybook.com/resources/how-to-track-traffic-from-aio-featured-snippets-paa-results-ga4/
Google. (n.d.). AI mode in Search . Google Blog. Retrieved from https://blog.google/products/search/ai-mode-search/
Schwartz, B. (2024). Google Ads AI Overviews Spotted . Search Engine Land. Retrieved from https://searchengineland.com/google-ads-ai-overviews-spotted-455889
Schwartz, B. (2024). Google Ads AI Overviews AI Mode Desktop . Search Engine Land. Retrieved from https://searchengineland.com/google-ads-ai-overviews-ai-mode-desktop-455733
