Organic CTR for queries where AI Overviews appear fell from 1.76% to 0.61% between June 2024 and September 2025, according to Seer Interactive's analysis of 3,119 queries across 42 client organizations. That's a 61% decline. Your rankings didn't move. Your content didn't get worse. The interface changed around you, and the measurement model most B2B teams rely on can't tell the difference.

The Three-Layer Problem

Citations, clicks, and conversions are three separate outcomes, and most dashboards treat them as if they're the same funnel. A page gets cited in an AI Overview. That citation doesn't reliably produce a visit. And a visit from an AI referral doesn't carry the same conversion profile as a branded search click. Pew Research Center found that in a panel study of 900 U.S. adults covering 68,879 Google searches, a link within an AI summary was clicked on about 1% of visits where a summary appeared, compared to 8% clicking a traditional result. When no AI summary showed up, traditional clicks hit 15%.

The citation layer is real, but the click layer is thin. If your content dashboard only reports sessions, you'll conclude "content stopped working" when the actual problem is that your measurement stopped covering the territory.

Why Citations Still Matter (Even Without the Click)

Seer Interactive's data, as reported by Search Engine Journal, showed that cited pages received roughly 120% more clicks per impression than uncited pages in AI Overview results. That's a relative lift, not an absolute traffic guarantee, and it's dataset-dependent. The directional signal is clear: being cited beats not being cited, even in a lower-click environment.

The harder question is what happens when the citation doesn't produce a click at all. A citation that puts your brand name in front of a buyer who later types your URL directly won't show up in your referral report. It will show up in your branded search trend, if you're looking. Attribution here is messy. That's not a reason to ignore the channel; it's a reason to triangulate.

What "Citable" Content Actually Looks Like

Content that AI systems can confidently reference tends to include clear facts, specific data, and expert context. That's a different brief than "rank for this keyword." A keyword-targeted page can rank without being citable, and a citable page can earn AI mentions without holding a top-10 position.

The practical move is to audit your content briefs for citability. Does the page contain a specific, defensible claim with a number attached? Does it answer a question a buyer would actually ask, in language that matches how they phrase it? Customer questions and objections (pulled from call transcripts, chat logs, support tickets) are a strong input here. One caveat: mirroring customer phrasing alone doesn't guarantee an AI citation. Credibility, brand recognition, and the structure of the answer all factor in. There's no formula that produces a citation with certainty.

Google Is Redesigning the Citation Layer in Real Time

On October 15, Google announced UI changes to AI Overviews and AI Mode: inline links placed higher, more prominent "Dive Deeper" links, tests of prominent official-site links, and expanded news carousels. The company also noted that Preferred Sources usage has grown more than sevenfold since the start of 2026. Its News AI pilot now includes more than 400 publishers.

None of these changes came with published click-through data. We don't know yet whether higher link placement translates into more referral traffic. What we do know is that Google is actively iterating on how sources appear inside AI answers, which means the citation layer isn't static. Teams that monitor their AI Overview presence will spot shifts before they show up in GA4.

The Measurement Model That Survives CTR Collapse

Ahrefs found that average CTR for the position-one result was 58% lower on keywords with an AI Overview compared to keywords without one, across 300,000 keywords in aggregated Search Console data. If your reporting model treats rank-one as a guaranteed traffic source, that model is already broken for a growing share of your keyword portfolio.

The replacement model separates three layers and measures each on its own terms:

Where This Leaves the Content Brief

The old brief said: target this keyword, match this intent, earn this rank. The new brief adds: make this page citable. Include a specific, defensible claim. Answer a real buyer question with enough structure and evidence that an AI system can reference it confidently. Then measure whether the citation shows up, whether branded demand moves, and whether the visitors who do arrive convert at a rate that justifies the effort.

That's three KPIs, not one. And the team that separates them first will stop misreading a measurement gap as a content failure.