Google Admits Search Console AI Search Reporting Falls Short — And Has No Fix Yet

Google's John Mueller confirmed on September 13, 2026 that Search Console's generative AI performance report is inadequate for measuring real AI search visibility — and that Google does not yet have a better solution.

The acknowledgment came in response to a Reddit thread that exposed fundamental flaws in how Search Console tracks impressions and positions within AI Overviews and AI Mode. For SEOs and site owners already navigating a rapidly shifting search landscape, the admission raises serious questions about whether the tools they rely on can keep pace with the technology driving modern search.

---

What the New AI Search Report Actually Shows

Google announced the Search Console generative AI performance report in June 2026. It began rolling out to a subset of websites before becoming globally accessible on August 31, 2026.

The report focuses primarily on impressions — tracking how many times a URL appears across AI search surfaces including AI Overviews and AI Mode. Critically, the data is a filtered view drawn from the regular search performance report rather than a standalone dataset. SEOs were warned not to add the two figures together to avoid double-counting.

On the surface, having any visibility into AI search performance seemed like progress. The reality, however, proved more complicated once practitioners began digging into the numbers. If you're evaluating Google's core tools for growing your business online, understanding the limitations of Search Console's AI reporting is now an essential part of that picture.

---

The Metrics Are Built on Outdated Assumptions

A Legacy Framework Struggling to Fit a New Reality

A Reddit user laid out the core problem in precise terms. The post argued that Search Console's AI Overviews metrics still follow "ten blue link legacy concepts" that do not reflect how AI search actually works — and the result is a reporting framework that can mislead site owners.

The Redditor broke down several specific issues that distort the data:

• An impression is counted when a link appears in an AI Overview that renders on the page — even if the user never scrolled down to see it
• Links hidden behind a "Show More" expansion are not counted until a user clicks to reveal them — meaning those impressions are understated rather than inflated
• The position metric records where the AI Overview block sits on the page — not where an individual link appears within that block
• Because the report is a filtered view of existing Web search data, SEOs who add both reports together are double-counting results

"Every link inside an AI Overview gets assigned the position of the AI Overview itself," the Redditor wrote, "so the average position you're looking at is the slot the AIO occupied on the page — not where your link sat among the others in it."

Mueller confirmed the assessment was accurate. "This is pretty much it," he wrote in response. "Position for these is hard to do in a way that makes it useful — so we're currently tracking it like we do for many search features (as a block) — and it's not separated out in the Gen-AI performance report."

Why This Matters More Than It Might Appear

The distortion runs in both directions. Impressions can be overstated when AI Overview blocks render below the fold without a user ever seeing them. At the same time, links that require a "Show More" click to reveal are actively undercounted. The result is a dataset that is simultaneously inflated in some dimensions and deflated in others — making it difficult to draw reliable conclusions from the numbers alone.

For site owners who have spent years calibrating decisions around Search Console data, this is not a minor inconvenience. It represents a fundamental mismatch between the metrics being reported and the user behaviour those metrics are supposed to reflect. Understanding how SEO analytics actually work and what the data means has never been more important than at this moment of transition.

---

Google Acknowledges the Paradigm Has Shifted — But Reporting Hasn't Caught Up

The Collapse of Position 1 Through 10

Mueller went further in his response by addressing the deeper structural tension between legacy search metrics and the reality of modern search engine results pages. The old position one through ten framework — once the universal currency of SEO performance — no longer maps cleanly onto how users experience search today.

"Search results pages have a lot of ways for users to interact nowadays," Mueller wrote. "The old 'position 1–10' is hard to map — or to make useful for site owners. If any of you have thoughts on what would be useful in terms of tracking position — I'd love to hear and am happy to discuss with the team."

The invitation for community input signals that Google is actively searching for a better model but has not yet identified one. For an industry that has spent decades optimising around position data — the absence of a reliable replacement metric leaves a meaningful gap.

The Measurement Problem Is Structural, Not Superficial

The situation reflects a broader reckoning across the technology world with AI-generated interfaces: the old maps no longer describe the new territory. As AI reshapes how search results are structured and consumed, the measurement tools built for a list-based world are struggling to adapt.

Mueller's comments confirm that Google is aware of the mismatch. Awareness, however, has not yet translated into a workable solution. The generative AI performance report remains in place as an imperfect instrument — relying on impression logic that can overstate visibility in some cases and understate it in others.

This challenge connects directly to a wider shift in how digital experiences are being designed and evaluated. As the nature of search changes, understanding search experience optimisation and what it means for your strategy is becoming as important as tracking rankings or impressions.

For now, SEOs are left interpreting data that Google itself acknowledges does not fully capture what matters: whether a real user saw a link — and where that link actually appeared within an AI-generated response.

What the Industry Should Watch Next

Google's open call for community input on improved position tracking is worth taking seriously. This is not a routine invitation — it reflects genuine uncertainty at the platform level about how to model user interaction with AI-generated results. SEOs who contribute structured, practical frameworks to that conversation may have a rare opportunity to influence how the next generation of Search Console reporting is designed.

In the meantime, the generative AI performance report should be treated as a directional instrument rather than a precise measurement tool. The underlying data is real — but the framework used to present it was built for a search environment that no longer exists in its original form. For further context on how Google approaches transparency around its own reporting limitations, the Google Search Central Blog remains the most reliable primary source for official updates as they are released.

---

How to Use This Information

• Treat Search Console's AI Overviews impression data as a directional signal rather than a precise measure of visibility — and avoid drawing firm conclusions from position metrics that reflect block placement rather than link placement
• Monitor "Show More" expansion behaviour separately when evaluating AI Overview performance, since those links are undercounted in current reporting until a user actively expands them
• Engage with Google's open call for input on improved position tracking — SEOs who contribute practical frameworks to that conversation may help shape the next generation of Search Console reporting