Google Admits Search Console Reporting for AI Search Is Inadequate

For years, we have relied on Google Search Console as the single source of truth for how our content performs. It is the primary tool we use to justify strategy shifts or prove the value of our work. But as search evolves into a generative experience, that source of truth is becoming blurred.

The problem is not just that the interface is changing, but that the underlying logic Google uses to measure success is fundamentally mismatched with how AI search actually works. When the tool we use to measure visibility is flawed, our decision making becomes a guessing game.

The rollout of Search Generative AI Performance Reports

In June 2026, Google introduced a new reporting layer within Search Console specifically for Search Generative AI. After a limited rollout to a small group of sites, these reports became available globally on August 31, 2026.

The primary goal of this report is to track impressions. It aims to show how often a specific URL appears within AI search surfaces, which includes AI Mode and AI Overviews. this is a filtered view. The data you see here is not additional traffic, but rather a subset of the data already present in your standard search performance reports.

From a practical standpoint, this means you are looking at a slice of your existing pie. The tradeoff here is clarity versus granularity. While it is helpful to see that a URL is appearing in AI surfaces, the report doesn't provide a distinct "AI only" traffic stream that is separate from the rest of the ecosystem. If you are making budget decisions based on these numbers, you must remember that these impressions are already baked into your general totals.

The disconnect highlighted by the community

The friction became public through a discussion on Reddit, where a user pointed out that the AI Overviews metrics are still tethered to the legacy concept of "ten blue links." The argument is that these metrics provide a distorted snapshot of actual performance.

The core of the issue lies in how an "impression" is counted. Under standard Google rules, an impression is logged if the item appears on the results page that is served, regardless of whether the user actually scrolls down to see it. In the context of an AI Overview, your link could be technically present in the generated response, but if the user never scrolls past the initial text, the impression is still counted. This creates an inflated sense of visibility.

Conversely, there is the "Show More" problem. If a link is tucked behind a "Show More" expansion, it is not counted as an impression until the user actually clicks to expand that section. This means your actual exposure could be higher than what the report suggests, but only if the user takes an extra action.

This creates a dangerous paradox for SEOs. You are simultaneously overstating your visibility (via non viewed impressions) and understating it (via hidden links). When analyzing these reports, you should question whether a spike in impressions actually correlates with a spike in clicks. If impressions are rising but clicks are flat, you are likely seeing the "phantom" impressions of the AI block rather than actual user engagement. A useful companion note is We Earned 1, because it looks at a nearby part of the same system.

Google's admission on reporting difficulties

John Mueller has since addressed these concerns, essentially agreeing with the assessment that the current reporting is inadequate. He acknowledged that it is difficult to present this data in a way that is truly useful for site owners.

Mueller pointed toward the help center documentation as the current source of truth, but admitted that the documentation is essentially a "giant doc" because the nuances are so complex. One of the most critical admissions is regarding "position."

Currently, Google tracks the position of the AI search feature as a single block. It does not track the specific position of your individual link within that AI block. For example, if the AI Overview is at the top of the page, the "position" is recorded for the block, not for where your URL sits inside that generated text.

This is a significant gap in intelligence. In traditional SEO, the difference between position 1 and position 3 is massive. In AI search, being the first cited source versus the fourth cited source likely has a similar impact on click through rates, yet Search Console cannot currently tell you which one you are. You are forced to treat the entire AI block as a binary: you are either in it or you are not.

Moving beyond the ten blue links paradigm

Mueller's commentary suggests that the industry needs to stop thinking in terms of a linear list of ten links. The modern search results page is a collection of diverse interactive elements, and trying to map that reality onto a 20 year old reporting framework is proving unsuccessful.

The reality is that Google is struggling to define what "position" even means in a generative context. While they have asked the community for input on what would be useful to track, the current state of the tool reveals a gap between Google's product evolution and its reporting evolution. They have changed how search works, but they haven't yet figured out how to tell us how we are performing within that new system.

The decision you need to make here is how much weight to give to GSC data for AI. If the tool is mapping a multi dimensional experience onto a linear scale, the data is a proxy, not a fact. I suggest shifting your focus toward conversion data and direct traffic patterns rather than relying solely on the "position" or "impression" metrics in the Gen AI report.

The reality of AI visibility tracking

When we strip away the technical jargon, the situation is simple: the tools we use to measure AI search are built for a version of the internet that no longer exists. The reliance on legacy concepts means that the data we see in Search Console is often a contradiction.

We see impressions counted for links that were never seen by human eyes, yet we miss impressions for links that were hidden behind a "Show More" button. We see the position of a feature block, but we remain blind to our actual rank within that block. This is not a failure of the SEO to understand the data, but a failure of the tool to provide accurate data.

The most important takeaway is that Google has admitted the reporting is inadequate. They do not currently have a better solution. Until a new framework for "generative visibility" is established, any AI specific data in Search Console should be viewed as a general trend indicator rather than a precise measurement of success. This connects with GSC’s New AI Overview Reporting when the same signal needs a clearer operating decision. The same pattern also shows up in Working Framework, where the practical question is how the signal becomes visible.

The risk of ignoring this is that you might optimize for "impressions" in AI Overviews, only to realize you are optimizing for a metric that doesn't actually represent a human seeing your brand. The goal should remain user intent and actual click throughs, as those are the only metrics that remain honest in an era of generative search.

Google Search Console Search Generative AI Performance Reports

The new Search Console report was announced in June 2026, initially rolling out to a "subset" of websites then fully accessible globally on August 31, 2026. The report primarily focuses on impressions, showing how many times a URL appears. The measurement question is whether this signal changes a decision, not whether it adds another number to a dashboard. Useful reporting connects visibility, engagement, and business outcomes without pretending every AI influenced journey will produce a clean click path.

The reporting question is whether this signal changes a decision. If it only creates another number in a dashboard, it adds noise. If it helps separate profile activity, website visits, calls, bookings, and direction requests, it can make local performance easier to understand.