How AI Overviews work, and why your traffic is dropping
- Jun 2
- 6 min read
Updated: Jun 9
AI Overviews now reach 2.5 billion users monthly and have fundamentally changed how organic traffic flows. When an AI Overview appears in search results, organic click-through rates drop by 58%. This post explains how AI Overviews are built, why businesses ranking on page one are still being skipped, and what to do about it.
Your organic traffic is dropping. Your rankings are holding. Your SEO agency has not flagged anything. So what is happening?
For a growing number of businesses, the answer is AI Overviews. Not a penalty. Not an algorithm update. A structural shift in how Google delivers answers, and who gets skipped in the process.
What is an AI Overview, and how does it actually work?
An AI Overview is a generated summary that appears at the top of Google's search results, above the organic links. Google assembles it from multiple sources, presents the answer directly on the page, and the user reads it without clicking through to any website.
It looks like a search result. It behaves like a final answer.
Understanding what happens under the hood matters, because it tells you exactly how your website can influence the output.
Not every query triggers one
AI Overviews are selective. Google evaluates whether generative AI genuinely adds value before showing one. Simple lookups and sensitive medical or financial queries are often excluded. The trigger is most likely for complex, research-driven, or conversational questions: the kind your potential customers are typing when they are actively evaluating options.
When it does trigger, here is the process
The process starts with the system expanding your question. Rather than running a single search for your exact phrase, it generates multiple related sub-queries simultaneously, building a custom pool of information specific to that question. It is not relying on pre-trained memory alone.
A large language model (LLM), the same type of technology that powers ChatGPT and Gemini, then drafts a natural language response based on what it retrieved.
But that draft does not go straight to your screen. This is where a technique called Retrieval-Augmented Generation, or RAG, comes in. Before the overview is shown, every factual claim in the draft is cross-referenced against live, indexed web content. If a claim cannot be verified against a reliable source, it gets dropped or rewritten. This is the mechanism that prevents the model from hallucinating, generating plausible-sounding but incorrect information.
The sources that successfully verified the AI's claims become the citations. The links inside an AI Overview are not random suggestions. They are the specific pages that proved the AI right.
What determines whether your content gets cited?
Two things, broadly. Whether your content can be extracted clearly, and whether it signals trust.
Google evaluates experience, expertise, authoritativeness, and trustworthiness (E-E-A-T) before using a source as a verification reference. Vague writing, anonymous pages, and content that cannot be cross-referenced against other reliable sources are more likely to be skipped entirely.
The practical implication: RAG means your website is not competing against a fixed database. It is competing to be the source that verifies the AI's claims in real time. That is a different problem from ranking, and most websites are not built for it.
How common are they?
AI Overviews now reach 2.5 billion users every month. At Google I/O in May 2026, Google announced the biggest redesign of Search in its 25-year history. The traditional results page is being replaced by AI-powered conversational answers. Links, Google's own product team acknowledged, are becoming an afterthought.
For queries in research-heavy categories such as health, finance, professional services, and software, the rate is even higher. The transition is not gradual. It is happening now, at scale, from the company that built the model your website was optimised for.
Why is organic traffic dropping?
Because the click is disappearing.
60% of all Google searches now end without a visit to any website. On mobile, that figure rises to 77%. When an AI Overview is present, organic click-through rates drop by 58%. In 83% of searches where an AI Overview appears, nobody clicks anything at all.
This is not a temporary fluctuation. Gartner projects that 50% or more of organic search traffic will be gone by 2028. The decline is not coming. It is already running.
Does ranking on page one protect you?
Less than you might expect.
73% of brands that rank on Google's first page receive zero mentions in AI-generated answers. Strong SEO and AI visibility are not the same thing, and optimising for one does not guarantee the other. The AI Overview pulls content from other sources and leaves them out entirely.
Does this mean traffic just matters less?
Not exactly. The volume story is changing, but there is an important counterpoint.
The visitors who do arrive through AI search are not the same visitors who arrived through traditional organic. In B2B and research-heavy sectors, AI search often sends fewer visits but materially higher-intent ones. Ahrefs reported a 23x higher conversion rate from AI-referred visitors on its own site.
Buyers who click through from an AI answer have already completed much of their evaluation inside the AI response. By the time they land on your site, they are not browsing. They are confirming.
What does this mean for your website?
For a full breakdown of how the buyer journey has changed and what your site needs to do differently, read our AEO and GEO explained guide. If you want to see how this plays out in numbers, the compressed funnel simulator models the impact for your sector.
The traffic drop is real. But the more important question is what happens to the traffic that does arrive, and whether your site is set up to close it.
What to do if AI Overviews are affecting your traffic
There are three layers to address, in order.
First, find out where you actually stand. Run your main service queries in ChatGPT, Perplexity, and Google as if you were a potential client. Note whether you appear in the generated answer, in the cited sources below it, or not at all. Those are three different problems with different fixes.
Second, fix your content extractability. The pages AI cites most reliably share one structural trait: they answer questions directly, with the answer in the first paragraph, not buried after brand context. Review your most important service pages and check whether a reader or an AI system can extract a clear answer from the opening 100 words.
Third, address the conversion layer separately. AI-referred traffic arrives later in the decision cycle and converts at a significantly higher rate than standard organic, but only when the landing page is built to receive it. Getting cited is step one. What the visitor finds when they click through is step two, and most businesses are only solving the first.
Frequently asked questions
What are Google AI Overviews?
Google AI Overviews are AI-generated summaries that appear at the top of search results for certain queries. They synthesise information from multiple sources and present a direct answer, reducing the need for users to click through to individual websites. They are currently most common for informational and research-stage queries.
Why does my organic traffic drop when AI Overviews appear for my keywords?
AI Overviews answer the query within the search results page, which reduces click-through rates for the organic results beneath them. If your traffic was driven by informational queries that AI Overviews now answer directly, those clicks may no longer happen, even if your ranking has not changed.
How does Google decide what to include in an AI Overview?
Google's AI Overviews draw on sources that it judges to be relevant, reliable, and well-structured. Factors that increase the likelihood of being cited include: clear, specific content that directly answers the query, consistent entity signals, established domain authority, and structured markup that helps Google parse the content accurately.
Can I get my content cited in Google AI Overviews?
There is no guaranteed route, but there are practices that increase the likelihood. Writing content that directly and specifically answers the queries your audience uses, structuring pages with clear headings and scannable answers, and building consistent authority signals across your domain all improve your chances of being selected as a source.
Is the traffic drop from AI Overviews permanent?
The nature of organic traffic is shifting rather than simply declining. Informational queries are increasingly answered within the results page. But decision-stage and transactional queries where users are choosing a provider or making a purchase still drive click-throughs. Businesses that focus their content on decision-stage queries are better positioned to maintain traffic quality even as overall volume changes.