Here’s a question that makes most agency clients uncomfortable: If your client ranks #1 on Google for their most important keyword, what are the odds the AI mentions them when a buyer asks about that category?

The answer, in early 2026, is roughly one in three — at best.

That number has collapsed fast. In July 2025, the overlap between Google’s top-10 organic rankings and AI Overview citations was around 76%. By February 2026, two separate analyses put it at 38% (Ahrefs) and 17% (BrightEdge). The relationship that once let agencies treat Google rankings as a proxy for AI visibility has broken down structurally — and most agencies haven’t told their clients yet.

This article explains exactly what’s happened, why it’s happening, and what it means for how you run AI visibility reporting.

The Numbers That Changed Everything

For about two years after AI Overviews launched, the conventional wisdom held firm: dominate traditional SEO, and AI visibility largely takes care of itself. The data seemed to support it. A 2024 study found that approximately 75% of AI Overview citations came from pages already ranking in the top 12 organic positions for the same query.

That correlation has now inverted.

A large-scale Ahrefs analysis published in early 2026 — covering 863,000 keywords and 4 million AI Overview URLs — found that only 38% of pages cited in Google AI Overviews also rank in the top 10 for the same query. That’s down from 76% just seven months earlier. A separate BrightEdge analysis using different methodology found the overlap even lower: approximately 17%.

The distribution of citations in the Ahrefs data is striking. Among AI Overview citations filtered to standard organic results:

Roughly two in three AI Overview citations now come from pages a user would never encounter on page one. The AI is drawing from a much larger pool than traditional search — and traditional SEO rankings are an increasingly unreliable predictor of whether a brand makes it into that pool.

Why This Is Happening: Query Fan-Out

The mechanism driving this shift is a process Google calls query fan-out.

When a user submits a search and an AI Overview is triggered, Google’s system doesn’t simply evaluate the top 10 results for that single query. It decomposes the original query into multiple related sub-queries — adjacent topics, related entities, alternative phrasings — and evaluates results across all of them. Pages that appear consistently and authoritatively across the full set of sub-queries become the cited sources.

The practical consequence: the source pool for any AI Overview is far larger than what ranks for the original keyword alone. A page ranking position 40 for a related sub-query may end up cited in an AI Overview triggered by the primary query, even though that page would never appear on the first SERP for the primary keyword.

Research confirms this. A Search Engine Land analysis of 10,000 keywords found that pages ranking across Google AI Overview fan-outs had a 161% higher citation probability than pages ranking only for the primary query. Topical breadth — covering a subject area comprehensively, not just targeting the head keyword — is what determines AI Overview inclusion in 2026.

This is why traditional SEO, which optimises heavily for keyword-by-keyword ranking, is structurally misaligned with how AI search selects sources.

The Signals AI Uses That SEO Doesn’t Track

Beyond query fan-out, AI platforms use a set of content selection signals that have no direct equivalent in traditional SEO.

Domain Authority correlation is falling. The Wellows analysis found that Domain Authority correlations with AI Overview inclusion dropped to r=0.18 in 2026 — a near-negligible relationship. High-DA sites still have an advantage, but it no longer dominates the selection process the way it dominates traditional rankings.

Structured data now carries enormous weight. SE Ranking research found that approximately 65% of pages cited by Google AI Mode include structured data markup. Wellows’ analysis associated structured data implementation with a 73% boost in AI Overview selection probability. This is a technical signal that traditional SEO monitors but rarely prioritises with the intensity AI visibility demands.

Third-party source presence matters more than brand-owned content. Research by Dr. Robert Li, drawing on SEMRush’s analysis of 2,500 prompts, found that the AI’s Stage 1 discovery process relies heavily on user-generated content platforms — Reddit, Quora, forums — rather than official brand websites. Only 12% of URLs cited by LLMs like ChatGPT and Perplexity appear in Google’s top 10 results, meaning AI platforms are actively drawing from sources that traditional SEO doesn’t even try to influence.

YouTube is now the most-cited domain in AI Overviews. Ahrefs’ analysis of 75,000 brands found that brand mentions on YouTube — in titles, transcripts, and descriptions — represent the strongest correlating factor with AI Overview visibility among all signals tested. YouTube grew its citation share by 34% in six months, with 18.2% of all out-of-top-100 AI citations now pointing to YouTube URLs.

None of these signals appear on a standard SEO dashboard.

What the Mention-Source Divide Means for Agencies

The research by Dr. Robert Li introduces a concept called the “mention-source divide” — the gap between brands that achieve high AI mention rates and brands that achieve both high mentions AND citations (with source links).

Only 3–27 brands per industry achieve both high mentions and citations simultaneously. The number varies by category: fashion shows minimal overlap (3 brands), while finance shows the highest (27). For most industries, a brand can be frequently mentioned in AI answers without ever being formally cited — meaning users hear the brand name but can’t click to verify it.

For agencies, this divide creates two separate optimisation tracks:

• Mention optimisation: Getting the brand name into AI responses through community presence, third-party mentions, and UGC signals
• Citation optimisation: Getting the brand formally cited with a source link through structured data, authoritative content, and E-E-A-T signals

Traditional SEO only touches the second track — and even then incompletely. AI visibility tracking is the only way to measure performance across both.

The Practical Implication for Your Reporting

If you’re currently reporting only on Google rankings and organic traffic, your clients are receiving an incomplete picture of their digital position. Here’s the specific gap:

A brand could hold the #1 Google ranking for its most important category keyword, earn 61% less organic CTR due to AI Overviews appearing above the fold, and simultaneously be absent from the AI Overview citing them — meaning competitors are receiving the AI’s implicit endorsement while your client’s #1 ranking generates almost no user interaction.

That’s not a hypothetical. It’s the default outcome for brands that optimise only for traditional search in 2026.

Specifically tracking AI citations — separate from Google rankings — is the only way to surface this gap, explain the traffic decline, and demonstrate the strategic adjustments needed. Platforms like PhantomRank track brand citation rates across AI platforms independently of Google rankings, making it possible to show clients exactly where they stand in the AI layer rather than inferring it from SERP positions.

Key Takeaways

• The overlap between Google’s top-10 organic rankings and AI Overview citations has fallen from 76% in July 2025 to between 17–38% in early 2026.
• Query fan-out means AI systems evaluate a much broader source pool than the top 10 results for any single keyword — pages outside top 100 now account for 36.7% of AI citations.
• AI selection signals — structured data, topical breadth, YouTube presence, third-party mentions — have no direct equivalent in traditional SEO tracking.
• The mention-source divide means brands can appear in AI responses without formal citations — and only AI visibility tracking can measure which side of that divide your client is on.
• Ranking #1 on Google is the foundation but no longer the ceiling. AI visibility requires a separate, dedicated tracking strategy.

For the mechanics behind why AI uses different sources than Google, read The Two-Stage Decision Architecture. To understand how to explain falling traffic to clients, see The Invisible Success Paradox.

Return to the AI Visibility Tracking Hub for the full framework.