In August 2026, Google confirmed a logging problem in Search Console's new Generative AI performance report. The issue was specific to reported data: impressions for generative AI in Search were understated for August 13 through August 17, and Google later said the missing data had been restored.
What exactly broke in Search Console?
Google's data-anomalies log says a logging error caused a decrease in impressions shown in the Generative AI performance report for Search from August 13 to August 17, 2026. On August 21, Google said the missing data had been restored and the Performance report showed complete metrics.
Did the bug mean Google Search traffic dropped?
No. This was a reporting issue, not a confirmed ranking or traffic problem. Search Engine Land reported Google's confirmation that the issue was in the Search Console data logging layer.
Why is this important now?
The incident matters because the Generative AI performance report is new. Google announced the report in June 2026 and said it was designed to show visibility in AI Overviews and AI Mode; on August 31, Google said the insights had rolled out to all websites worldwide.
What does the Generative AI report actually measure?
The report focuses on impressions from supported generative AI features on Google Search. You can group the data by dimensions such as page, country, device and date. It is best treated as a visibility dataset rather than a normal keyword-ranking report.
What should you do when the report looks wrong?
First check Google's Search Console data-anomalies page. Then compare the affected dates with your normal Performance report and, where useful, Analytics. Do not label a sudden graph change as an SEO loss until you have ruled out a reporting anomaly. Google explicitly recommends checking its anomaly documentation when investigating unusual Performance report changes.
How should SEOs report AI-search visibility?
Keep the source metric clear. For example, say “Generative AI impressions in Search Console” rather than “AI clicks” when the report does not provide clicks. Also record the report date range and any known anomalies in the same dashboard or notes.
What is the bigger lesson?
First-party measurement is useful, but it is still a reporting system with definitions, aggregation rules and occasional logging errors. The strongest AI-search reporting setup combines Search Console, Analytics, page-level diagnostics and a clear record of Google data anomalies.
Related ToolBoxKart guides
For the main reporting workflow, read the Search Console AI Performance Report Guide and the Google AI Search Visibility Report. To act on what the data shows, use the guide to optimizing pages for Google AI Search and check regional differences with Google's regional Search experience guide.