TL;DR — Key Takeaways
- +38% organic clicks when Google disables AI Overviews — from Google's own randomized controlled experiment (reported by Search Engine Journal, May 9, 2026).
- Chartbeat documents -60% search traffic for small publishers over 24 months, -47% for mid-size publishers.
- Google added 5 new link types to AI Overviews on May 8 — still without providing click data in Search Console.
- As of May 11, 2026, zero separate GSC data exists for AI Overview clicks vs. traditional organic results.
- What to do: GEO (Generative Engine Optimization) becomes the priority — become the source AI cites, not just a ranked result.
The experiment Google probably shouldn't have disclosed
A number shook the global SEO community on May 9, 2026: in a randomized controlled experiment run by Google itself, disabling AI Overviews from search results caused organic clicks to jump 38%. The finding was reported by Search Engine Journal, citing data presented in a regulatory hearing.
This isn't a third-party study. This isn't an extrapolation. This is Google, in a controlled experiment, documenting the impact of its own features on publisher traffic — and then refusing to publish that data in Search Console.
Context matters: this data surfaced in US legal proceedings where Google must justify its practices regarding publisher traffic. The company had previously argued AI Overviews don't significantly harm click-through rates. These numbers contradict that position.
What it means in practice: If you rely on informational evergreen content (how-to guides, tutorials, "what is X" articles), AI Overviews are directly cannibalizing your clicks. Google answers the question in the search results — users no longer need to click through to your article.
Chartbeat: 2 years of data, -60% for small publishers
Chartbeat, the real-time analytics platform used by hundreds of web publishers, released its 24-month audience data (2024–2026). The results are stark:
- Small publishers (organic traffic <500K visits/month): -60% search-driven traffic
- Mid-size publishers (500K–5M visits/month): -47% search-driven traffic
- The decline accelerated after each AI Overview expansion (May 2025, October 2025, March 2026)
For businesses that rely on SEO content to generate leads, this signal is critical. A blog that drove 10,000 organic visits per month two years ago may now be generating 4,000 — with traditional analytics tools unable to clearly explain why.
Is your organic traffic plateauing or declining? Cicero Studio performs a complete GEO + SEO audit to identify whether AI Overviews are cannibalizing your content — and which articles to reprioritize.
Get my free article →Free article →5 new link types in AI Overviews — still no click data
On May 8, 2026, Google announced the addition of 5 new link features inside AI Overviews:
- Inline contextual links — placed directly next to the text they support (no longer grouped at the end)
- "Explore new angles" section — related articles displayed beneath the AI synthesis
- Primary source links — more links to the original studies and source documents
- Multi-source enriched answers — visible aggregation of multiple publishers for the same question
- Topic follow-up links — redirections toward deeper content on a subject
In theory, these changes increase publisher visibility within AI Overviews. In practice, there is still no way to measure whether these new links generate clicks — or how many.
Google has shifted its messaging on AI Overview impact four times since launch, according to a count by Search Engine Land. The latest official version: "AI Overviews send more traffic to publishers than before." The Chartbeat data and the randomized experiment suggest a more nuanced reality.
The GSC blind spot: measuring the invisible
This is the core structural problem: Google Search Console does not separate click data for AI Overviews from traditional organic results. As of May 11, 2026, this data simply doesn't exist in standard GSC reports.
What you can observe indirectly:
- A rise in impressions without a proportional rise in clicks (your content is "cited" but not clicked)
- A declining CTR on queries where your position is stable
- A maintained average position while organic traffic drops
If you observe this pattern in your GSC data, AI Overviews are likely cannibalizing your content.
For more on the broader shift in AI search visibility measurement, see our analysis of how OpenAI's move into advertising changes the GEO equation.
What this means for your content strategy
Given this data, three strategic repositionments are essential:
1. Become the cited source, not just the ranked result
Google AI Overviews operate on the same principle as ChatGPT, Perplexity, and Gemini: they synthesize the most trustworthy and citable sources. GEO (Generative Engine Optimization) is the discipline that targets exactly this: making your content structured, factual, and citable by AI systems.
This means: named data with sources, clear definitions, a Q&A structure, and explicit entities (people, brands, places). It's the strategic evolution beyond traditional on-page SEO.
2. Identify at-risk vs. protected content
AI Overviews don't appear uniformly. According to Ahrefs data cited by Search Engine Journal, appearance rates vary significantly by category:
- Health / Science: ~45% of queries trigger an AI Overview
- Finance / Legal: ~35%
- Technology: ~30%
- Breaking news: ~15% (too volatile for AI synthesis)
- Local / transactional queries: <10%
If your content is heavily informational on "universal" topics (health, law, general technology), your exposure is maximum. Content with proprietary data, first-hand testimonials, or niche angles is less exposed.
3. Invest in timely content and fresh data
AI Overviews cannot synthesize information that AI systems don't yet know. An article published today about an announcement from today has a direct organic visibility window before AI syntheses integrate it. This is part of why breaking news traffic surged while evergreen traffic collapsed.
Cicero Studio — GEO audit + editorial production + automated semantic linking. Agency-quality work, software-grade productivity.
See plans →Limits of this analysis
What this article doesn't cover
- Results vary by sector: Chartbeat data is aggregated — your specific vertical may be more or less exposed to AI Overviews.
- Google's randomized experiment was not published in a peer-reviewed academic journal. It emerged in a legal context with methodological conditions not independently verifiable.
- Partial correction is possible: being cited in an AI Overview maintains brand awareness even without a click. This indirect impact is not quantified here.
- E-commerce and local sites are less impacted than pure informational content publishers — the click mechanics are different.
Frequently asked questions
Yes, documented by Google itself. A randomized controlled experiment shows a 38% organic click spike when AI Overviews are disabled. Chartbeat documents -60% search traffic for small publishers over 2 years. Both sources converge on the cannibalisation effect.
No. As of May 11, 2026, Google Search Console provides no separate click data for AI Overviews vs. organic results. Indirect signals (falling CTR with stable impressions and positions) can suggest impact, but there is no precise measurement available.
Content with proprietary data, named sources, first-hand expertise (strong E-E-A-T), and a unique editorial angle. Breaking news also resists well: AI Overviews appear on only ~15% of news queries, compared to ~45% for health and science topics.
Being cited in an AI Overview maintains brand visibility even without a direct click. The GEO goal is to become the source AI cites — requiring citable content, named data, and a clear factual structure. This is what Cicero Studio helps businesses build, as a full-service GEO + SEO agency.
Growth and SEO & GEO content strategy specialist, I founded Cicéro to help businesses stay visible as AI search transforms the rules of organic traffic. Every article we produce is designed to be cited by AI, not just indexed by Google. Agency-quality work, software-grade productivity.
LinkedIn