If you’ve been tracking your SEO visibility and feeling good about the trend line, I have a number that should worry you. Across roughly 2,600 companies, the correlation between a company’s organic search gains and its AI Overview citation gains over the same four weeks is effectively zero. Not low. Zero.
That means your ranking is telling you almost nothing about whether AI is citing you. You could be climbing in Google search results and disappearing from ChatGPT and AI Overviews in the exact same month. Same company. Same content. Two completely different verdicts.
That’s not a glitch. It’s a pattern. And once you see what’s driving it, you’ll understand why some of your competitors are winning a game you didn’t know had two boards.
This comes from Kevin Indig’s Growth Intelligence Brief — if you don’t know his work, Growth Memo is one of the sharpest, most rigorously sourced reads in SEO right now, and I’m crediting him directly because the data underneath this is his team’s, not mine. They track something called the Search Signals Index — thousands of companies, across two dozen-plus industries, measured on two separate dimensions: how visible they are in organic search, and how often AI systems mention them.
Most of the industry still treats those two numbers as basically the same thing, moving together. This report proves they’ve split apart — and the way they split apart tells you exactly what kind of company wins each game.
Here’s the mechanism. Google’s core organic ranking system rewards the page that collects and compares — the aggregator, the roundup, the directory, the review hub — because a human reading a page of ten blue links still has to make a decision. The aggregator helps them decide. That’s valuable real estate, and Google pays for it with rankings.
AI Overviews work backwards. The AI has already made the decision for the reader. It doesn’t need the page that compares five options — it needs the actual source underneath the comparison. So it skips the collector and cites the thing being collected.
Same query. Same web. Two different winners, because the two systems are optimizing for two different moments — one for a reader who still needs to browse, one for a reader who already got the answer.
The data backs this up at real scale. Companies classified as aggregators — 168 of them in the index — gained over 11% in organic visibility in four weeks, and lost over 5% of their AI Overview mentions in that same window. Companies classified as integrators — primary sources, 170 of them — gained on both sides at once: visibility up, AI mentions up.
You can see it in the names, too. Reddit, LinkedIn, Facebook, and Instagram all posted double-digit organic visibility gains in the same four weeks they lost meaningful ground in AI Overview citations. LinkedIn alone jumped by over 40% in visibility, while its AI mentions remained flat. Meanwhile, Barnes & Noble lost more than half its visibility in the same window — mid-tier retailers gave up the shelf space the biggest platforms just reclaimed.
And there’s a seasonal tell buried in here worth sitting with. Home improvement retailers — Wayfair, Home Depot, Lowe’s, Ace Hardware — saw AI mentions jump 25 to 50% in four weeks while their organic visibility barely moved. Summer project demand is appearing in AI shopping results before it appears in search rankings. The AI layer is pricing seasonal intent faster than the organic index does.
So here’s what this means for you, and I’d rather you sit with the question than take a checklist from me.
Look at your own site honestly. Are you the collector, or are you the source? If your value is comparing, curating, or aggregating what other people built, you should expect your AI citations to erode even while your rankings look fine. That’s not a bug; it’s not fixable with better formatting. It’s the structure of what you built.
Here’s what it looks like when a company sits on the other side of that line. Burbank Dental Lab used to publish static PDF success guides — good content that had already driven $2.4 million in revenue on its own. But a PDF is a dead end for an AI system. It can’t crawl inside it. It can’t extract one passage and cite it. It just sits there, readable by humans, invisible to everything else.
So we are currently building it into an interactive 3D platform that teaches complicated implant procedures, with an AI guide embedded directly on the page. Now the same expertise isn’t just readable — it’s structured to be extracted. And in its infancy, it’s already working: we’ve confirmed, via Ahrefs, a live backlink to that platform from CAD-Ray, a company in the same clinical ecosystem. Burbank Dental Lab isn’t hoping to get chosen from a list. They built the thing that gets pointed to.
If you already are a primary source — you have the original data, the real client outcomes, the testing nobody else did — none of that matters if it’s locked in a format the AI can’t extract and cite. A PDF nobody can parse doesn’t count. A page built to be read, not lifted, doesn’t count.
There’s one more number from this report I keep coming back to. AI Overviews and AI Mode — two different AI surfaces from the same company — moved in opposite directions for the same brands, in the same four weeks. They correlate with each other at just 0.30. If you’re only tracking one “AI visibility” score, you’re already missing half the picture, just as organic-only tracking missed the aggregator collapse.
So don’t ask me for your AI visibility number. Ask yourself which one of these two rankers actually rewards what you built. If the honest answer is neither — that’s not a tactics problem. That’s the thing to fix first.
If you’ve been tracking your SEO visibility and feeling good about the trend line, I have a number that should worry you. Across roughly 2,600 companies, the correlation between a company’s organic search gains and its AI Overview citation gains over the same four weeks is effectively zero. Not low. Zero.
That means your ranking is telling you almost nothing about whether AI is citing you. You could be climbing in Google search results and disappearing from ChatGPT and AI Overviews in the exact same month. Same company. Same content. Two completely different verdicts.
That’s not a glitch. It’s a pattern. And once you see what’s driving it, you’ll understand why some of your competitors are winning a game you didn’t know had two boards.
Two numbers that used to move together just split apart — and nobody’s tracking the gap.
where this comes from
This comes from Kevin Indig’s Growth Intelligence Brief — one of the sharpest, most rigorously sourced reads in SEO right now, and I’m crediting him directly because the data underneath this is his team’s, not mine. They track the Search Signals Index: thousands of companies, across two dozen-plus industries, measured on two separate dimensions — how visible they are in organic search, and how often AI systems mention them.
Most of the industry still treats those two numbers as basically the same thing, moving together. This report proves they’ve split apart — and the way they split apart tells you exactly what kind of company wins each game.
TWO Rankers, two moments
Google’s core organic ranking system rewards the page that collects and compares — the aggregator, the roundup, the directory, the review hub — because a human reading a page of ten blue links still has to make a decision. The aggregator helps them decide. That’s valuable real estate, and Google pays for it with rankings.
AI Overviews work backward. The AI has already made the decision for the reader. It doesn’t need the page that compares five options — it needs the actual source underneath the comparison. So it skips the collector and cites the thing being collected.
Same query. Same web. Two different winners — because the two systems are optimizing for two different moments.
What The Numbers Show
The data backs this up at real scale. Companies classified as aggregators — 168 of them in the index — gained over 11% in organic visibility in four weeks, and lost over 5% of their AI Overview mentions in that same window. Companies classified as integrators — primary sources, 170 of them — gained on both sides at once: visibility up, AI mentions up.
You can see it in the names, too. Reddit, LinkedIn, Facebook, and Instagram all posted double-digit organic visibility gains in the same four weeks they lost meaningful ground in AI Overview citations. LinkedIn alone jumped over 40% in visibility while its AI mentions remained flat. Meanwhile, Barnes & Noble lost more than half its visibility in the same window — mid-tier retailers gave up the shelf space the biggest platforms just reclaimed.
And there’s a seasonal tell buried in here worth sitting with. Home improvement retailers — Wayfair, Home Depot, Lowe’s, Ace Hardware — saw AI mentions jump 25 to 50% in four weeks while their organic visibility barely moved. Summer project demand is appearing in AI shopping answers before it shows up in search rankings.
THE QUESTION TO ASK ABOUT YOUR OWN SITE
Look at your own site honestly. Are you the collector, or are you the source? If your value is comparing, curating, or aggregating what other people built, you should expect your AI citations to erode even while your rankings look fine. That’s not a bug you can fix with better formatting. It’s the structure of what you built.
Here’s what it looks like when a company sits on the other side of that line. Burbank Dental Lab used to publish static PDF success guides — good content, and it had already driven $2.4 million in revenue on its own. But a PDF is a dead end for an AI system. It can’t crawl inside it. It can’t extract one passage and cite it.
So we are rebuilding it as an interactive 3D platform teaching implant procedures, with an AI guide embedded directly on the page. Now the same expertise isn’t just readable — it’s structured to be extracted. And in its infancy, it’s already working: we’ve confirmed, via Ahrefs, a live backlink into that platform from CAD-Ray, a company inside the same clinical ecosystem. Burbank Dental Lab isn’t hoping to get chosen from a list. They built the thing that gets pointed to.
If you already are a primary source — you have the original data, the real client outcomes, the testing nobody else did — none of that matters if it’s locked in a format the AI can’t extract and cite. A PDF nobody can parse doesn’t count. A page built to be read, not lifted, doesn’t count.
Half of your AI visibility might be invisible to whatever you’re currently measuring.
THe NUMBER THAT STAYS WITH ME
AI Overviews and AI Mode — two different AI surfaces from the same company — moved in opposite directions for the same brands, in the same four weeks. They correlate with each other at just 0.30. If you’re only tracking one “AI visibility” score, you’re already missing half the picture, just as organic-only tracking missed the aggregator collapse.
So don’t ask for your AI visibility number. Ask yourself which one of these two rankers actually rewards what you built. If the honest answer is neither — that’s not a tactics problem. That’s the thing to fix first.
So don’t ask for your AI visibility number. Ask yourself which one of these two rankers actually rewards what you built. If the honest answer is neither — that’s not a tactics problem. That’s the thing to fix first.
So ask yourself this week: are you actually building the foundation — or are you just buying a new coat of paint and calling it a strategy?
The proof isn’t a theory. It’s 338 companies moving in opposite directions on cue.
COMPLETE THE FORM TO
BOOK A STRATEGY CALL
"*" indicates required fields
CHAPTERS:
COMPLETE THE FORM TO
BOOK A STRATEGY CALL
"*" indicates required fields
Subscribe and Share – WE APPRECIATE YOUR SUPPORT
