The number that stopped me was sitting inside a Search Console segment I had filtered to AI Overviews only. My client's page was ranking number two for their primary target query. Solid, stable, two years of work behind that position. The AI Overview for the same query was citing three sources. None of them were my client's page. One of them ranked eighth.
I had already been skeptical of the new consultancy category that had started appearing: AI Overview optimization sold as a distinct discipline sitting alongside traditional SEO. The pitch usually involved proprietary frameworks and new deliverables dressed in terminology that made the work sound more novel than it was. What the 52 percent figure from Google's own data confirmed is that the category is real in outcome but not in method. Google is building AI Overviews primarily from content it already trusts at the ranking level. You do not optimize for AI Overviews separately from ranking. Ranking is the prerequisite.
But the data also confirmed something the pitch decks were not going to say directly. Being in the top 10 makes you eligible. It does not make you chosen.
The page ranking eighth that appeared in the AI Overview instead of my client's second-place page was not better written. It was not more authoritative by any backlink or entity metric I could measure. What it had was a structural habit the AI citation layer appears to reward: direct, declarative answers positioned at the opening of each content section, before any context or qualification. Not a featured snippet format with a definition box. Not a numbered how-to structure. Just prose that front-loaded the answer and then expanded, rather than building toward the answer through a paragraph.
Image credit: Screenshot from "SEO in 2026: How I'd Rank in Google in the AI Era" by Ahrefs on YouTube (https://www.youtube.com/watch?v=tiW6xRYSXmM).
My client's page did the opposite. It was written the way thoughtful long-form content gets written: context first, answer second, nuance throughout. That structure works well for a human reader who has committed to reading. It works less well for a system extracting a concise answer to cite in a summary. The page that ranked lower but answered faster was the one that got pulled.
I rebuilt the top three sections of my client's page around that observation. Not a full rewrite. Just a structural inversion of how each section opened. Rankings did not change. The page held at two. Four weeks later it appeared in the AI Overview for the primary query. Also for three related queries it had not previously been cited for. The only variable that changed was the sentence-level structure at the top of each section.
That result is encouraging, but it comes with a direct warning attached. The 52 percent figure means the floor for AI Overview visibility is a top-10 ranking. If you are not there, the structural work is irrelevant. There is a version of this conversation happening in a lot of agencies right now where teams are spending energy on AI Overview citation tactics for pages that rank fourteenth and wondering why nothing is moving. The answer is that the citation layer is downstream of the ranking. Fix the ranking first.
The part that has not been honestly reckoned with yet is what this means for businesses that have never been able to break into the top 10 for their most valuable queries. AI Overviews do not create a secondary path to visibility for low-authority sites. They consolidate visibility further toward what is already winning. A page that has never ranked in the top 10 for a given query is not going to appear in the AI Overview for that query regardless of how well its content is structured. For small businesses watching AI Overviews appear above the organic results on queries they were already struggling to compete on, the experience of losing click share to a format that draws from the same top 10 they were already excluded from is not a new problem. It is the existing problem displayed in a more visible way.
What has changed is the stakes attached to breaking into that top 10. A year ago, ranking tenth meant you received a fraction of the traffic the first three positions got and that was the full cost of the position. Now ranking tenth means you are in the pool that gets cited in AI Overviews. Ranking eleventh means you are not. The gap between tenth and eleventh is larger than it has ever been, and most of the industry is still pricing and scoping SEO work as if the reward for incremental ranking improvement follows a linear curve.
Getting a page from rank four to rank two used to be worth something. Getting a page from rank twelve to rank nine is now worth considerably more than it looks on the surface, and the work required to do it is exactly the work it has always been.
Waleed Qamar holds a BSc in Computer Science from Purdue University and has spent the years since turning that technical foundation into something the curriculum never covered: figuring out why websites rank, why they fall, and why most businesses never find out until it is too late.
Pakistan-born and based between the United States and South Asia, he has managed search visibility for e-commerce stores, local service businesses, and SaaS startups across two continents. He started in SEO when guest posting still worked, survived the Penguin update, and has rebuilt client sites from scratch after algorithm hits more than once.
He has watched good businesses get sold packages that looked like progress and delivered nothing lasting. He has also seen the right approach quietly double a site’s traffic without a single press release about it.
His writing on SEO By Highsoftware99 covers Google algorithm updates, autocomplete optimization, semantic SEO structure, and the widening gap between what agencies promise and what Google actually rewards in 2026.
He knows what a traffic cliff looks like in Search Console on the morning you discover it.
