Your AI visibility number just broke in half

I was on a call last week with a marketing director who pulled up her team's blended 'AI visibility' dashboard, one clean line trending up and to the right, ready to call it a win for the quarter. I asked her three questions: which platform, which engine, and up compared to what. She didn't have separate answers for any of them. Neither did her dashboard. That's the whole problem in miniature, and this month there's data showing exactly what that one blended number is starting to cost the people who trust it.

SEO visibility and AI citation just stopped moving together

For most of the last two years, the assumption behind AI search visibility tracking was simple, even if nobody wrote it down: AI citation and organic SEO visibility drift together. When a brand's rankings climbed, its AI mentions tended to climb too, roughly on the same timeline, for the same pages. Track one number, call it 'AI visibility,' and treat it as a decent stand-in for both — the same convenient shortcut behind the claim that AI visibility can substitute for the Google clicks it's replacing.

Kevin Indig's Growth Intelligence Brief #21, published on his Growth Memo newsletter on July 10, 2026, is the first clean data I've seen that the assumption just broke. Looking at the five largest platforms by AI mention volume, he found that June 2026 was the first month organic SEO visibility and AI Overview citations diverged for all five at once. Not diverged the way monthly noise usually looks. Diverged like a brand climbing in one column while falling in the other, on the same pages, in the same month.

“SEO visibility and AI citation used to drift together for the mega-platforms. In June, they came apart.”

Put numbers on it and the picture gets sharper. Some of the biggest names on the web moved a lot in June, and the direction of that move had almost nothing to do with whether their AI mention volume moved the same way.

LinkedIn didn't just grow, it grew during the same stretch that the aggregate AI mention count barely moved system-wide, a detail that matters for the next section. Barnes & Noble's organic visibility index fell from 24.76 to 11.04, more than half gone in a single month, with Alibaba and Shutterfly posting comparable drops. None of that swings evenly with AI citation volume for the same brands. A platform can be rising in organic search and losing ground in AI answers, or the reverse, and a reporting setup that only tracks one combined 'visibility' score will show you neither story, just a misleading average sitting somewhere in the middle. That's a bad place to make a budget decision from, and it's exactly where a mention-rate dashboard built on one blended number will quietly steer you wrong.

The plateau that was hiding a redistribution

Here's the part that should worry anyone reporting a single 'AI mentions' figure to leadership. Indig also tracked total AI mention volume across the platforms in his dataset, and for eight straight weeks it sat nearly flat: roughly 6.0 million mentions a week, barely moving. If you were only watching that top-line number, June looked uneventful. Nothing to report. Steady as she goes. That flat line was doing a good job of hiding an actual redistribution happening underneath it.

AI Overviews mentions slipped 3.2% over the same period. Add those three moves together and the total comes out looking almost unchanged, which is exactly the trick. A flat aggregate isn't evidence that nothing happened. Sometimes it's evidence that two or three things happened in opposite directions and canceled each other out on the one chart anybody bothered to look at. If your reporting and analytics setup only produces that one combined number, an eight-week plateau like this one would have sailed by without a single follow-up question.

AI Overviews vs AI Mode: what real AI search visibility tracking would show

Break the plateau apart by engine and the story turns into something you'd actually want to act on. AI Mode mentions rose 22.5%. ChatGPT mentions fell 28.4%. AI Overviews mentions slipped a smaller 3.2%. That's not three engines drifting near each other. That's one engine gaining real ground, one losing a meaningful chunk of its share, and one holding roughly steady, all inside a period where the combined number told you approximately nothing.

This is the piece generative engine optimization reporting keeps getting wrong, in my experience: teams build a GEO analytics view with one AI citation rate at the top, then wonder why it doesn't explain a client's traffic shifts from one quarter to the next. If ChatGPT is losing ground to AI Mode for a given brand, that brand's overall citation count can hold steady while its actual exposure to ChatGPT's users quietly craters. How we track AI citations at the URL level is built around exactly this problem — a citation that shows up fine in the aggregate but has moved almost entirely from one engine to another isn't the same win it looks like on a summary slide.

What home improvement proves about AI search visibility tracking

The home improvement category is the cleanest argument for splitting the numbers, because it broke the overall pattern entirely. Wayfair's AI mentions grew 48.1%. Ace Hardware's grew 45.8%. Across the category, Indig reported 25-50% AI-mention growth in June, while organic visibility for those same brands stayed close to flat.

Indig's read on it is sharp: AI surfaces react to seasonal demand shifts faster than organic search does. Someone typing a question about deck stain or patio furniture into an AI engine in June gets a fresher answer, faster, than the same query would have surfaced through classic organic rankings a year earlier. The AI side of the ledger moved because the season moved. The organic side hadn't caught up yet, or didn't need to.

That gap is only visible if you're tracking AI mention volume and organic SEO visibility as two different lines for the same brand. Collapse them into one blended score and Wayfair's June looks unremarkable. It wasn't. It was a genuinely strong month, in a channel a combined metric was never built to show you. If you run SEO and GEO for anyone with real seasonality, from home goods to a large enterprise retailer's category pages, this is the exact blind spot that costs you credit for a win that already happened.

Split the number before it lies to you again

None of this requires a research team. It requires refusing to let one chart stand in for three different questions. Pull your organic visibility trend for your top pages. Pull your total AI mention or citation count for the same set, separately. Then break that count out by engine, at minimum ChatGPT, AI Overviews, and AI Mode, and watch for a month where the total stays flat while the engines underneath it don't. If you find that pattern, and after June 2026 you probably will, that's the month your reporting has to catch, not the month after it shows up in a client's numbers and nobody can explain why.

It's worth saying plainly what this isn't. It isn't a case for abandoning a single summary number in the boardroom. Leadership still wants one line on one slide, and that's fine, as long as somebody on the team can defend the three lines underneath it when a client asks why the summary moved. The mistake isn't having a headline metric. It's letting the headline metric be the only metric anyone ever pulls up, so a month like June slides through unnoticed until the gap between what the dashboard says and what actually happened gets too wide to explain away.

The marketing director on that call wasn't wrong to want one number. Everybody wants one number; it's easier to defend in a meeting than three. But this month is the proof that the easy number and the true number just split apart, at least for the biggest platforms on the web, and averaging them back together doesn't make the split go away. It just makes sure you're the last person in the room to notice it.

See where you are cited today

A free snapshot audit of your rankings and AI citations before we ever talk.