{"id":22223,"date":"2025-11-13T11:14:41","date_gmt":"2025-11-13T10:14:41","guid":{"rendered":"https:\/\/inboundcph.dk\/?p=22223"},"modified":"2026-11-09T23:20:48","modified_gmt":"2026-11-09T22:20:48","slug":"google-ai-overviews","status":"publish","type":"post","link":"https:\/\/inboundcph.dk\/en\/google-ai-overviews\/","title":{"rendered":"Google AI Overviews: What Are They, and How Are You Cited?"},"content":{"rendered":"
\n

What are Google AI Overviews?<\/h2>\n

Google AI Overviews are the AI-generated answers that Google displays at the top of the search results, above the regular links.<\/strong> The answer is generated by Google's language model, Gemini. The model gathers information from multiple websites and displays links to the sources it has used. In Danish, this feature is called \"AI-oversigt.\".<\/p>\n

Google launched AI Overviews in the U.S. in May 2024. In Denmark, the rollout was initially delayed because \"Danish Media in 2024\" was rejected<\/a>, that Google was allowed to use content from Danish media outlets in its responses. AI Overviews Go to Danish search results in May 2025<\/a>.<\/p>\n

For a business, AI Overviews mean there are two positions to compete for in the same search. One is the ranking in the list of links. The other is the spot as a source within the AI answer itself. The two positions are related, but they are not the same. At InboundCPH, our metrics show pages that rank highly organically without being mentioned in the AI answer, and pages further down the list that Google cites anyway. Here, we explain how Google selects its sources and what you can specifically do to get cited.<\/p>\n

\n
\n
58%<\/span><\/div>\n
Lower click-through rate for the top result when Google displays an AI Overview (Ahrefs, 2026)<\/div>\n<\/div>\n
\n
8%<\/span><\/div>\n
clicks on a regular result when an AI response is displayed (Pew, 2025)<\/div>\n<\/div>\n
\n
56\/56<\/span><\/div>\n
AI-SEO searches tracked by InboundCPH trigger an AI Overview (Nov. 2026)<\/div>\n<\/div>\n<\/div>\n

When does Google display an AI Overview?<\/h2>\n

Google displays an AI Overview when the system determines that a comprehensive answer adds value beyond the usual links.<\/strong> Google itself states this in its Documentation on AI features in Google Search<\/a>. AI Overviews therefore do not appear in nearly all searches. Long questions with an informational intent most often trigger an AI response. Searches containing a brand name do so less frequently, and the same applies to local searches and news.<\/p>\n

An analysis of 146 million search results, as reported by Search Engine Journal in November 2025, showed that AI Overviews appeared in 57.9% of question-based searches. On average, across all searches, the percentage was 20.5%.<\/p>\n

\n\n\n\n\n\n\n\n\n\n\n\n
Search type<\/th>\nAI Ad Mix Overview<\/th>\n<\/tr>\n<\/thead>\n
Questions (\u201cHow\u2026?\u201d)<\/td>\n57,9 %<\/td>\n<\/tr>\n
Long searches (7+ words)<\/td>\n46,4 %<\/td>\n<\/tr>\n
Searches without a brand name<\/td>\n24,9 %<\/td>\n<\/tr>\n
Average for all searches<\/td>\n20,5 %<\/td>\n<\/tr>\n
Searches by brand name<\/td>\n13,1 %<\/td>\n<\/tr>\n
Local searches<\/td>\n7,9 %<\/td>\n<\/tr>\n
News Searches<\/td>\n6,3 %<\/td>\n<\/tr>\n<\/tbody>\n<\/table>
Source: Search Engine Journal, November 2025 (146 million search results, U.S.).<\/figcaption><\/figure>\n

When the buyer faces a complex decision, the percentage is significantly higher. InboundCPH tracks 56 Danish searches for AI-SEO and SEO on our own AI visibility platform, and in November 2026, all 56 trigger an AI Overview. These are questions such as \u201cIs AI SEO worth the money?\u201d and \u201cHow much does an SEO campaign with AI cost?\u201d These are precisely the considerations a decision-maker goes through before contacting a vendor. If you\u2019re selling something that requires careful consideration, you should therefore expect AI Overviews to appear in the search results that matter most to your sales pipeline.<\/p>\n

How does Google select the sources for an AI Overview?<\/h2>\n

Google selects the sources for an AI Overview from among pages that are indexed and eligible to appear as a snippet in Google Search. There are no additional technical requirements.<\/strong> This is stated in Google\u2019s own guidelines. To construct the answer, Google uses a technique called query fan-out. The model sends out multiple related searches across subtopics and compiles the answer from the pages that best answer each sub-question. That\u2019s why a page might be cited for a single paragraph, even if it isn\u2019t the best result for the main search.<\/p>\n

Organic rankings still matter a great deal. seoClarity analyzed 432,000 keywords in October 2025 and found that 99% of AI Overviews draw their sources from the top ten organic results. However, a top-10 ranking is a ticket to the door, not a guarantee. When Google needs to quote a snippet, the model selects text that provides a clear answer and can stand on its own. A paragraph that doesn\u2019t answer the question until the fourth sentence loses out to one that does so in the first.<\/p>\n

The third signal is mentions outside your own site. Ahrefs analyzed 75,000 brands and found that online mentions of a brand are about three times more strongly correlated with visibility in AI answers than backlinks (correlation of 0.66 versus 0.22). Mentions in industry media and on comparison sites are therefore an integral part of working with AI Overviews\u2014not a side note. We go into more depth on Google\u2019s own recommendations in our Guide to Google's AI Optimization<\/a>.<\/p>\n

In Google's own words<\/span><\/p>\n

Google states in its 2026 documentation that no special optimizations are required to appear in AI Overviews. Nor are new machine-readable files or special schema markup needed. Our interpretation is that the classic SEO principles remain unchanged, but that competition is fiercer because only a few sources are included in the result.<\/p>\n<\/div>\n

How much do AI Overviews cost per click?<\/h2>\n

AI Overviews result in significantly fewer clicks on the top results, but pages cited in the response perform better than those that are not.<\/strong> In May 2026, Ahrefs compiled the results of 13 studies and concluded that The click-through rate for the top result drops by 58%<\/a>, when an AI Overview is displayed. The Pew Research Center conducted a survey in July 2025<\/a>, that users click on a regular search result in 8% of visits with an AI response, compared to 15% without one.<\/p>\n

There\u2019s another side to the story. In November 2025, Search Engine Land reported on a study by Seer Interactive that examined exactly that. Brands mentioned in an AI Overview receive 35% more organic clicks and 91% more paid clicks than brands not mentioned in the same search results. The loss, therefore, hits those left out of the results the hardest.<\/p>\n

The clicks aren\u2019t the whole story, either. The AI response builds the buyer\u2019s shortlist, and much of the decision-making takes place within the response itself. If a buyer has seen your name in an AI Overview, she\u2019ll often Google the company name right afterward, and that visit is recorded as brand or direct traffic. At InboundCPH, we say that you\u2019re measuring the shadow and overlooking what casts it. Low AI traffic in Google Analytics isn\u2019t the same as low AI impact.<\/p>\n

How to Become a Source in Google's AI Overviews<\/h2>\n

You can become a source in AI Overviews by having the page that answers the question most clearly hosted on a domain that Google already trusts.<\/strong> At InboundCPH, we work according to our 5-layer method, which ranges from the entity through the questions, answers, and authority to the measurement. You can read more about the foundation in What is AI-SEO?<\/a>. The six steps below are the practical version for a single page or topic.<\/p>\n

\n
1<\/div>\n
\n

Find the questions that trigger AI responses<\/h3>\n

Search for your most important questions in an incognito window, and note where Google displays an AI Overview and who is cited. Write the questions as a customer would say them out loud, not as keywords. Also consider follow-up questions, since Google\u2019s query fan-out looks for answers to sub-questions.<\/p>\n<\/div>\n<\/div>\n

\n
2<\/div>\n
\n

Provide one quotable answer for each question<\/h3>\n

Formulate the headline as a question, and have the first sentence answer it in no more than 40 words. The paragraph should be able to stand alone, so refer to the topic by name instead of using \u201cit\u201d or \u201cas mentioned.\u201d This is the part of AI-SEO that is also called GEO (Generative Engine Optimization)<\/a>.<\/p>\n<\/div>\n<\/div>\n

\n
3<\/div>\n
\n

Enter numbers and names into the text<\/h3>\n

AI models prefer to cite a sentence that includes numbers, units, years, and sources rather than a sentence containing words like \u201cmany\u201d or \u201cmost.\u201d Name specific products and methods. The model cannot use numbers that appear only in a graphic, so include them in the body text.<\/p>\n<\/div>\n<\/div>\n

\n
4<\/div>\n
\n

Remove the technical barriers<\/h3>\n

The page must be indexed and must not have \"nosnippet\" or \"max-snippet:0,\" because then Google cannot display a snippet from it. Make sure that robots.txt and your CDN (e.g., Cloudflare) aren\u2019t blocking Googlebot. Important content must appear as text and not be hidden in images or tabs that require JavaScript. Structured data must match the visible text.<\/p>\n<\/div>\n<\/div>\n

\n
5<\/div>\n
\n

Build buzz outside your own site<\/h3>\n

Get your name and your data included in the sources that AI models draw from, especially industry media and comparison sites. At InboundCPH, we use the platform\u2019s data on which Danish media outlets the AI engines actually cite to select placements.<\/p>\n<\/div>\n<\/div>\n

\n
6<\/div>\n
\n

Measure citations, not just rankings<\/h3>\n

Follow a set list of search queries each month, and track whether the AI response appears and whether you are the source or are mentioned. Without this tracking, you won\u2019t know if the changes are working.<\/p>\n<\/div>\n<\/div>\n

Why isn't a page in the top 10 cited?<\/h2>\n

A page in the top 10 is typically not cited because Google ranks the page as a whole, whereas AI Overviews cite specific, clear answers.<\/strong> Google can easily deem a long guide relevant enough to rank it, while at the same time finding a better-formulated answer to the sub-questions on a competitor\u2019s site. The difference rarely lies in the topic; it lies in how the text is written.<\/p>\n

Consider a hypothetical example. An auditing firm has a detailed page about interim CFOs that ranks sixth in the search results. The page opens with three paragraphs about the firm\u2019s history, and the price for an interim CFO is listed first in a table at the bottom. A competitor ranked 9th has a short section titled \u201cHow much does an interim CFO cost?\u201d, and the first sentence provides a daily rate range and a year. Google cites the competitor because the answer can be pulled directly into the AI Overview. The auditing firm doesn\u2019t need a new page. It needs to move the answer up and rewrite it so that it can stand on its own.<\/p>\n

\n
Ranked, but not cited<\/span><\/p>\n

This text is written to be read from beginning to end<\/h3>\n