A share of your audience's searches already ends without a single click: the person reads the AI summary and decides right there. If the sources cited in that block are your competitors, you lose the sale without even showing up in the report.
In this guide, I'll show you when AI Overviews appear, what changes in CTR, and what to do with your content and code to become a cited source.
What Google AI Overviews are
AI Overviews are answers generated by Artificial Intelligence that Google displays at the top of the results page, above the traditional links. The block is assembled by Gemini, which summarizes the topic and cites the sources used in side cards. It shows up mainly in informational, comparative, and question based searches, and has been running in Brazil in Portuguese since 2024.
In practice, Google stopped merely listing pages and started answering. And the question that matters for anyone running marketing is simple: is your company among the cited sources, or did it get left out of the summary?
How Google builds the generated answer
The process has a name: query fan-out. When you search for something, Google breaks your question into several silent sub-searches, fires them all at once, and retrieves specific passages from different pages for each part of the answer.
Gemini then stitches those passages into a single text and links the sources. Note the detail: Google doesn't take the whole page, it takes passages. Whoever writes direct, quotable paragraphs has a better shot at being clipped. That's exactly the logic a GEO strategy works with.
Difference between AI Overviews, featured snippet, and AI Mode
Many people confuse the three blocks, and the distinction changes how you optimize:
- Featured snippet: a literal excerpt from ONE page, copied and pasted at the top. It's been around since 2014.
- AI Overviews: new text, generated by AI from SEVERAL sources, with citations alongside. It coexists with the regular results.
- AI Mode: a separate tab, chat style, where the conversation continues with follow-up questions. The traditional SERP disappears.
The featured snippet rewards whoever answers best in a single paragraph. AI Overviews reward whoever is a trustworthy source across several parts of the topic. Similar games, different rules.
Why appearing in AI Overviews changed the organic traffic game
If your dashboard shows stable rankings and falling sessions, you've probably already felt the effect. This scenario has a name: zero-click search. The user asks, Gemini answers right there, and a good share of people never reach your site.
The session metric alone hides the problem. You're still in position 3, Search Console shows impressions rising, but the click vanishes. Anyone looking only at traffic concludes the content got worse. In practice, the storefront changed.
The real impact on organic position CTR
The public numbers are harsh. A 2025 Pew Research Center study showed that the click-through rate on organic results drops from 15% to 8% when there's an AI Overview on the page. Ahrefs measured a 34.5% drop in first position CTR on searches with a generated answer.
Translating that to your funnel: the same keyword, at the same ranking, can deliver a third fewer visits. And that's without a single technical mistake on your side.
The other side of the coin matters: whoever is cited as a source inside the block inherits the attention the blue links lost. The competition shifted from position to citation.
Which types of search trigger more generated answers
The impact varies by intent. The most affected categories today:
- Informational: "what is", "how it works", "how much does it cost on average"
- Comparative: "X or Y", "best tools for"
- Health, finance, and legal questions, where Google prioritizes synthesis with sources
Brand searches and direct transactional queries remain less affected. If your business depends on informational top of funnel, you're at the epicenter.
How to prepare your site to be cited in AI Overviews
Gemini doesn't cite whoever has the prettiest text. It cites whoever delivers the answer that's easiest to extract. That's what we work on.
Content structure the AI can extract
Start every page with an answer block: a paragraph of 40 to 60 words that answers the main question directly, right below the H1. Then go deeper.
Write headings the way people search ("how much does it cost", "how it works", "what's the difference"). Use numbered lists for processes and tables for comparisons. AI loves predictable structure.
This work of becoming a citable source has its own discipline, GEO, and it starts in the content architecture, not in the technology.
Structured data and technical signals
Mark up your pages with schema: Article, FAQPage, Organization, Product when it makes sense. Structured data works like a caption so the machine understands what each passage means.
In robots.txt, check whether you're accidentally blocking Googlebot or Google-Extended. A slow page or one whose content depends on heavy JavaScript is also out of the race.
Authority and mentions outside your domain
E-E-A-T carries weight here. An identified author with a real bio, proprietary data, cited sources. Plus mentions of your brand in media outlets and third-party sites, because Gemini cross-checks what you say with what the market says about you.
Mistakes that kill your eligibility
The most common ones I find in audits: thin, mass-produced content, pages with no direct answer (all preamble), broken schema or schema inconsistent with the visible text, and cannibalization, with five URLs competing for the same question. Fix that before expecting a citation.
How to measure whether your site appears in AI Overviews
Measurement closes the loop. Without it, you don't know whether the content and schema work turned into a citation or just into cost.
Manual tracking and monitoring tools
Start simple. List 20 to 50 priority queries, the ones that actually generate pipeline, and search each one weekly in an incognito window. Record it in a spreadsheet: did an AI Overview appear? Was your brand cited? Which competitor made it into the card?
Manual works to get started, but it doesn't scale. Semrush and Ahrefs already flag which keywords trigger AI Overviews in their position tracking. Here at Netlinks we use App Netlinks, which monitors citations across dozens of AIs live and cross-references that data with the structured data audit of each page. That way you can see what changed technically when a citation appears or disappears.
How to separate AI traffic in GA4 and Search Console
In Search Console, clicks coming from AI Overviews are lumped into the "Web" type, with no dedicated filter. The indirect signal is clear: stable or rising impressions with falling CTR on informational queries.
In GA4, create a segment with a referral regex covering chatgpt.com, perplexity.ai, gemini.google.com, and copilot.microsoft.com. That isolates the external AIs.
And here comes the point I repeat most in meetings: compare revenue and conversion rate for that segment, never sessions. A visitor arriving from a generated answer has already read the summary and comes in decided. Measuring that gain is the foundation of any well executed GEO strategy.
Where to start before the next report
That drop in sessions with stable rankings that opened this article now has a name and a cause: Gemini answers at the top of the page and cites whoever prepared for it. The problem is no longer invisible, which changes everything.
Three things to take away:
- AI Overviews reward extractable answers. A 40 to 60 word answer block below the H1, question format headings, and structured data aligned with the content.
- Clicks became rarer, but the citation became the new top of the SERP. Whoever shows up in the card enters the buyer's consideration set ahead of the competitor.
- Without weekly measurement of your priority queries, you don't know whether you're inside the block or outside it. A simple spreadsheet handles the start.
Tomorrow morning, run the test: take your 10 most important searches, run each in an incognito window, and write down who Gemini cites. If your brand isn't there, you've just mapped where the pipeline is leaking.
At Netlinks, this work runs inside the 8 pillars of the NETLINKS Method, with live citation monitoring across dozens of AIs in App Netlinks, 24 hours a day. Since 2017, more than 200 companies have gone through this process, with over R$ 100 million in sales generated for clients.
Being the answer became the new first place.
If the incognito window test revealed a gap, Talk to an Expert and get a diagnosis of your situation in AI Overviews, with no strings attached.
Frequently asked questions
What are Google AI Overviews?
AI Overviews are answers generated by Artificial Intelligence that Google shows at the top of the results page, before the traditional links. The block is assembled by Gemini, which summarizes the subject and displays the consulted sources in cards next to the text. It already works in Brazil in Portuguese and appears most often in informational and comparative searches and question based queries.
Do AI Overviews appear in every search?
No. They mainly show up in searches with informational intent, such as "how it works", "what's the difference", and "is it worth it". Brand searches, direct navigation, and very short transactional terms still show classic results more often. In practice, the more a search looks like a question, the higher the chance Gemini will build a summary at the top.
Do AI Overviews reduce my site's traffic?
On informational queries, yes, CTR tends to fall, because some users read the summary and don't click anything. That's what's known as zero-click search. On the other hand, whoever is cited as a source gains visibility and receives more qualified clicks, from people who read the answer and wanted to go deeper. The real risk is being outside the block while a competitor shows up inside it.
How do I get my site cited in AI Overviews?
Deliver the answer ready for extraction. Open every page with a 40 to 60 word paragraph answering the main question right below the H1, write headings in the format people actually search with, and use verifiable data with sources. Round it out with structured data (FAQPage, Article, Product, or LocalBusiness schema, as applicable) so Google understands the page context effortlessly.
What's the difference between AI Overviews and GEO?
AI Overviews are a specific Google feature, the answer block generated by Gemini on the results page. GEO (Generative Engine Optimization) is the whole discipline: optimizing the brand to be cited by any generative engine, including ChatGPT, Perplexity, Claude, and AI Overviews themselves. In other words, appearing in Google's block is one of the outcomes of a well executed GEO strategy.
How do I know whether my brand appears in AI Overviews?
Build a list of 20 to 50 queries that generate business, search each one weekly in an incognito window, and record it in a spreadsheet: did the block appear, was your brand cited, which competitor made it in. In Search Console, compare impressions and clicks for those queries over time, because stable impressions with falling clicks usually indicate an AI Overview sitting in front of your result.
Does schema markup help you appear in AI Overviews?
It helps, quite a lot. Structured data doesn't guarantee a citation, but it makes the page easier for Google to read, allowing it to identify what's a question, an answer, a product, an author, and a review. FAQPage, HowTo, Article, and Organization schemas are the most useful in this context. Think of them as a technical caption for your content: the less ambiguity for the machine, the greater the chance of becoming a source.
