The Search Console Generative AI Performance Report Explained

#01What does the report cover?

Google's generative AI Search report help describes impressions for your pages when they appear as links in AI Overviews and AI Mode. Google announced the reports in June 2026, with data available at hourly, daily, weekly and monthly granularity. A separate Discover generative AI report covers generative AI features in Discover.

#02How is this different from the Performance report?

Google's AI features documentation says AI Overviews and AI Mode are counted in the Web search type of the Performance report, using the same rules as other results. So the Performance report already includes AI feature impressions and clicks, mixed in with everything else. The generative AI report isolates the impressions, so you can see which pages are used as sources.

#03How should you read multimodal data?

Google added Web: multimodal as a search type in September 2026 for searches that start with images, such as Lens, Circle to Search and uploaded images. In the generative AI report, splitting text-based and multimodal shows whether your product photos or diagrams lead people to AI answers that use your pages. Lesson 15 covers multimodal reporting in full.

Our view: treat generative AI impressions as a source-share signal. Rising impressions on pages you care about mean Google trusts them as grounding. It's a better signal for content quality than rankings on a single keyword, but it says nothing about revenue.

#04Reading the generative AI report alongside Performance

The generative AI report has impressions only. To understand what AI features do to your traffic, put it next to the Performance report for the same pages and dates.

These are hypotheses to test, not conclusions. Search Console doesn't show clicks from AI features separately, so you can't measure AI Overview CTR directly.

#05Troubleshooting

#06What the report doesn't tell you

• Clicks. Impressions only. Clicks from AI features are in the Performance report's Web totals, unlabelled.
• Queries. There's no query dimension, so you can't see which prompts triggered your page.
• Position inside the answer. You don't know if you were the first or the fifth source.
• Other assistants. ChatGPT, Claude, Perplexity and Copilot are invisible here (lesson 20).

#07A repeatable analysis, from export to decision

1. Export generative AI impressions by page for the last 3 months, Weekly.
2. Export Performance by page (Web search type) for the same dates, with clicks, impressions, CTR and position.
3. Join on URL in a sheet. Remember the generative AI report uses the canonical URL, so normalise trailing slashes and parameters before matching.
4. Add a section column (blog, glossary, product, help) using the URL path.
5. Pivot by section: generative AI impressions, Web clicks and Web CTR. Look at direction over the period, not the ratio between reports.
6. Pick actions: protect pages with high AI exposure and stable clicks; improve pages that should be sources but aren't; accept that some definition pages will lose clicks.

#08Edge cases

• Canonical consolidation. If several URLs share a canonical, their AI impressions appear under the canonical only. A page that "has no AI impressions" may be credited to its canonical.
• Hourly view. Useful on launch days or after changing the generative AI control, but preliminary data changes, so don't report from it.
• Small sites. You may see very few rows, or none. That's normal for low-volume properties; check again monthly.
• Country splits. AI Overviews and AI Mode availability varies by country and language, so compare countries only with care.

#09Worked example: Ledgerline's AI source check

Ledgerline is a fictional accounts-payable software company. Details and numbers are illustrative.

Ledgerline's SEO lead exported the generative AI pages list. Five glossary pages accounted for most AI feature impressions, but the pricing comparison page, a priority, didn't appear. Performance showed the glossary pages' CTR easing while impressions held, consistent with AI Overviews answering definitions. The team decided not to fight that: they added a short "how Ledgerline handles this" section with a link to the product page on each glossary entry, and annotated the change to check in two months.

Walkthrough

1. Open Performance > Search generative AI (the label may differ slightly as the report evolves).
2. Set the date range to the last 3 months and Weekly granularity.
3. Open the Pages tab and sort by impressions. Export the list.
4. Switch the search type between Web: text-based and Web: multimodal to see whether image-led searches matter for you.
5. Check the Discover generative AI report if your site gets Discover traffic.
6. Open the standard Performance report, filter to your top 5 AI-surfaced pages and compare clicks and CTR over the same period.
7. Annotate any date you changed the generative AI control or published major updates.

#10Try it yourself

Build an AI-feature exposure sheet:

1. Export the top 50 pages by generative AI impressions for the last 3 months.
2. Add each page's Performance clicks and CTR for the same period.
3. Label each page with one reading from the table above.
4. Choose two pages to watch monthly and one to improve.

#11Quick quiz

1. Does the generative AI report show clicks?

Answer: No. It reports impressions only.

2. Which date timezone does it use?

Answer: Pacific Time, like other Search Console reports.

3. Are Search Labs experiments included?

Answer: No. They are excluded.

4. Are AI Overview clicks counted anywhere?

Answer: Yes, inside the Web search type of the Performance report, but not labelled separately.

#12What should you check before the next lesson?

• You found the generative AI report in your property.
• You exported your top AI-surfaced pages.
• You compared them with Performance clicks for the same period.
• Read What the Search Console AI Overviews data shows for an earlier analysis and forecasting traffic loss from AI Overviews.

Next: The Search generative AI control. Or return to the Search Console for Marketers track.

#13Common questions

Can I see which queries triggered an AI Overview citation? No. The generative AI report has no query dimension.

Does AI Mode count in the normal Performance report too? Yes. Google counts AI features within the Web search type of the Performance report.

Why do my generative AI impressions differ from Performance impressions? The reports count differently and the generative AI report aggregates by canonical page, so compare trends rather than totals.