Search Console: bug in AI performance reports

Google Search Console recently introduced performance reports for generative AI search—and on June 24, 2026, it confirmed the first documented bug in this reporting area. Google points to a data logging bug affecting Discover as well as Generative AI in Discover performance reports. For SEO teams trying to measure visibility in AI-powered surfaces, this is an important signal: the new metrics are not yet fully reliable.

With the Generative AI performance reports, Google is responding to the growing importance of AI Overviews, AI citations, and related search formats. Publishers now receive structured data on how content appears in generative contexts for the first time. In parallel, Google continues to provide classic search performance data for web, images, and video. The current bug, however, specifically affects the intersection of Discover and generative AI—an area that is especially relevant for news and content publishers.

What Google says about the bug

According to official documentation, this is a data logging error in Search Console. The affected views are Discover performance reports and the Generative AI in Discover report. Google confirms that the data captured in these views is currently not being logged correctly. That means clicks, impressions, and derived metrics may appear distorted or incomplete in the affected reports.

The timing is notable because the Generative AI reports were rolled out only recently. Many teams have just integrated the new filters into their monthly reporting routines. An early bug underscores that the data foundation for GEO and AI search analytics is still in an introduction phase—and that caution is warranted when interpreting trends.

Which reports are affected—and which are not

The documented bug is limited to Discover and Generative AI in Discover. Classic web search reports, image and video performance, and other established Search Console views are not covered by this notice. Teams that evaluate only organic web search data do not need to question their existing dashboards across the board.

However, anyone specifically measuring visibility in Discover or AI elements within Discover should treat the affected reports with skepticism for now. It becomes especially critical when decisions about content budgets, editorial priorities, or GEO strategy are based solely on these still faulty numbers.

Discover and generative AI in context

Google Discover delivers personalized content recommendations outside classic search queries. With generative AI elements in Discover, recommendation logic and AI-generated summaries blend together. For publishers, that means additional visibility channels—and new metrics. Exactly this interface is currently unreliable because of the logging bug.

Practical recommendations for SEO and analytics teams

Until Google fixes the bug, a cautious approach to the affected reports is advisable. Comparisons with prior periods should be drawn only with clear reservations. Where possible, teams should use supplementary signals: server log analysis, referrer data from analytics, manual SERP checks, and qualitative observations of AI citations in live search.

Internally, stakeholders should be informed that Generative AI reports in Discover cannot currently serve as a reliable basis for final decisions. A brief note in monthly SEO reporting prevents misinterpretation and protects against premature optimization measures based on faulty numbers.

• Mark affected reports as provisional for now.
• Do not base major budget or strategy decisions solely on Discover AI data.
• Use supplementary data sources in parallel.
• Check Google documentation and status notices regularly.
• Revalidate historical data after the bug fix.

Implications for GEO and AI search monitoring

Generative engine optimization depends on measurable visibility in AI answers and related surfaces. As long as central Search Console reports log incorrectly, part of GEO steering remains in the dark. That is not an argument against the new reports—on the contrary, they show that Google is expanding measurability. At the same time, the bug makes clear that early product generations in Search Console have not always been stable right away.

Teams with established technical SEO should include the bug in their quality assurance. Check whether exported CSV data, API connections, or Looker Studio dashboards include the affected metrics. A short audit of the data pipeline saves correction effort later, once Google fixes the logging issues.

What to expect after the fix

Once Google corrects the logging error, the affected reports can gradually flow back into standard dashboards. It makes sense to reconcile the numbers again with the period immediately before and after the fix. In some cases, there may be backfills or corrections when previously mislogged events are processed retroactively.

For the SEO community, the incident remains a lesson: test new Search Console features early, but establish them as binding KPI sources only after stabilization. The documented bug on June 24 affects only one segment—but it reminds teams that AI search reporting is still young and continuous quality control remains essential.

Communication with management and editorial teams

In larger organizations, technical notices in Search Console are not enough on their own. Analytics owners should briefly explain why Discover AI metrics may be excluded from management slides or editorial OKRs for now. A factual framing prevents missing trends from being misread as content decline—or short-term spikes from being celebrated as sustainable success.

In parallel, it is worth updating internal playbooks for AI search monitoring. Document which data sources count as primary and which as secondary signals during the bug. That keeps the team actionable without ignoring the new reporting layer entirely.