GSC now reveals which pages surface in AI Overviews

Google has, until now, kept marketers in the dark about whether their pages were being cited inside AI Overviews. Search Engine Journal reports that has changed: Google Search Console now exposes which URLs appear in AI Overviews, AI Mode, and Discover, giving site owners their first official read on visibility inside generative answers.

This is the most consequential GSC update since the rollout of Core Web Vitals, and not because the data is rich. It is because the data exists at all.

For two years, the only way to know whether a brand was cited in an AI Overview was to prompt Google repeatedly and screenshot the result, or pay a third-party monitoring tool to do the same at scale. Both methods are noisy. Both miss the personalised variants Google serves to different users. Marie Haynes, writing in Search Engine Journal, notes that the new reporting collapses AI Overview, AI Mode and Discover impressions into the existing Performance report, which means the same interface that has governed SEO reporting for a decade now also governs LLM visibility reporting. That matters more than the feature itself.

What the data actually shows, and what it hides

The reporting confirms whether a page appeared, not where it ranked inside the answer, not which query variant triggered it, and not whether the citation drove a click. Google still aggregates AI Overview impressions with standard search impressions in the top-line numbers, which means anyone wanting a clean read has to filter by search appearance. Click-through rates from AI Overviews are, predictably, lower than from blue links. Haynes points out that pages cited in Overviews often see impressions rise and clicks fall, the now-familiar pattern across generative search.

The useful signal is comparative. Which of your pages get pulled into Overviews, and which do not? The answer reveals what Google's retrieval layer treats as citation-worthy on your domain, which is almost never the same set of pages that rank highest in classic results. Listicles, definitional explainers, and pages with clear entity markup tend to surface. Thought leadership PDFs and gated reports do not.

The reporting gap closes for everyone at once

For B2B brands, the strategic value sits in benchmarking. A multilateral publishing climate guidance, an industrial group explaining a new standard, a bank producing market commentary: all three have spent the past eighteen months guessing whether their content reaches the audiences that now ask Gemini or AI Mode instead of typing keywords. They can now stop guessing. The pages Google chooses to cite are the pages worth doubling down on; the pages it ignores are candidates for restructuring or retirement.

There is a quieter implication for agencies and in-house teams that have sold AI visibility audits as a premium service. The moat shrinks when the underlying data becomes free. Third-party tools, Profound, Peec, Athena and others, will still matter for cross-model coverage (ChatGPT, Perplexity, Claude), but the Google-specific portion of their pitch just got commoditised. Expect pricing pressure within the quarter.

What it changes for content strategy

The honest read is that GSC's new view will expose how thin most enterprise content libraries are when measured against generative retrieval. Financial services brands that publish quarterly outlooks will discover that their fund commentary gets cited; their corporate ESG narrative does not. UN agencies will find their factsheets surface; their executive speeches do not. Industrials will see product specification pages and standards documentation outperform brand campaigns by an order of magnitude.

This is not a failure of those brands. It is a clarification of what generative search actually values: structured, specific, declarative content with clear entities and clean provenance. The teams that act on the new GSC data will rebuild content roadmaps around the citation patterns it reveals. The teams that treat it as another dashboard will keep producing prose that no model retrieves.

One caveat worth holding. Google controls both the index and the report. Anyone relying solely on GSC for AI visibility data is trusting the referee to keep its own scorecard. Pair it with independent monitoring across the other models, because the brands winning citations in Gemini are frequently invisible in ChatGPT, and vice versa. The window Google just opened looks onto Google's house only.