Search Console’s Generative AI performance report shows how often your pages appear in Google AI Overviews and AI Mode. It reports impressions by page, country, device and date, and nothing else: no clicks, no click-through rate, no position and no queries. Used well, it answers one question clients keep asking: “Are we showing up in Google’s AI answers?” Used badly, it produces big numbers that mean very little.
This guide covers what the report measures, the traps to avoid, and a five-slide structure that turns it into decisions. It’s written for agencies building client reports and for in-house marketers who need to explain AI visibility to leadership.
Key takeaways
– Google launched the report in June 2026 and, per its help page, rolled it out to all sites worldwide on August 31, 2026.
– It shows impressions only for AI Overviews and AI Mode, by page, country, device and date. Data starts May 18, 2026.
– AI impressions are already counted in the standard Performance report’s Web totals. Never add the two together.
– Pair it with the Performance report and GA4 to see what happens after the impression.
– The best slide isn’t the total. It’s which pages Google trusts in AI answers, and what you’ll do about that.
What does the Generative AI performance report show?
It shows how many times links to your site appeared in Google’s generative AI features on Search. According to Google’s documentation, it covers:
- Features: AI Overviews and AI Mode, combined into one number.
- Metric: Impressions, meaning how many times a link to your site was shown in one of those features.
- Dimensions: Pages, countries, devices and dates.
There’s a separate report for AI features in Discover.
What doesn’t the report show?
This matters more than what it does show, because clients will assume the gaps aren’t there.
| Missing | Why it matters | Workaround |
|---|---|---|
| Clicks and CTR | You can’t see if AI visibility sends visits | Check the same pages’ clicks in the Performance report and GA4 landing-page data |
| Queries | You don’t know which searches triggered the citation | Join the AI report export to Performance data by URL to see likely queries |
| Position | No idea where in the AI answer you appeared | Spot-check key queries by hand |
| AI Overviews vs. AI Mode split | One combined number | None yet. Note it in the deck |
| History before May 18, 2026 | No year-over-year comparison until mid-2027 | Use month-over-month and page-level trends |
There’s also an export limit. Independent testing published in September 2026 found the Search Console API doesn’t yet accept a generative AI search type, so the data has to be exported from the interface by hand.
Why can’t you add AI impressions to regular impressions?
Because they’re already in there. Google says AI feature data continues to be included in the Performance report under the Web search type. The Generative AI report is a separate view of part of that data, not extra visibility.
Adding the two inflates your numbers and will eventually be caught by a sharp client. Put a one-line footnote on any slide that shows both: “AI impressions are a subset of total Web impressions.”
How should you pull the data each month?
A repeatable 20-minute routine:
- Export the Pages tab from the Generative AI report for the last full month and the month before.
- Export the Performance report (Web, Pages tab) for the same dates, with clicks, impressions, CTR and position.
- Join them by URL in a spreadsheet. You now have, for each page, AI impressions next to total impressions, clicks and CTR.
- Add GA4 data for those landing pages: sessions, engagement and key events (leads, sales).
- Flag three groups:
– High AI impressions, healthy clicks and conversions: protect and update.
– High AI impressions, falling clicks: the AI answer may be satisfying the search. Check whether the page offers a reason to click.
– Important pages with no AI impressions: candidates for content upgrades.
What should the client deck look like?
Five slides. Executives read the first one; managers read the rest.
Slide 1: The headline
One sentence and one chart. Example wording:
“Google showed our pages in AI Overviews and AI Mode [X] times last month, up/down [X]% from the month before. Most of that visibility came from [X] pages. Leads from organic search were [steady/up/down].”
Slide 2: Which pages Google trusts in AI answers
A table of the top 10 pages by AI impressions, with their clicks, CTR and conversions from the joined data. This is the most useful slide in the deck, because it shows which content Google is pulling from and whether that content is earning anything.
Slide 3: Gaps
Important pages (service pages, key guides, product categories) with few or no AI impressions. Each gets a one-line diagnosis: thin answer, outdated, not indexed, weak internal links, or simply a topic where AI answers don’t appear.
Slide 4: What we’re doing next
Three to five actions tied to slides 2 and 3, each with an owner and a date. For example: “Rewrite the opening of [page] so it answers [question] in the first two sentences,” or “Add [original data / expert quote] to [page].”
Slide 5: Caveats
Short and honest:
- Impressions only; no clicks or queries in this report.
- AI Overviews and AI Mode are combined.
- Already included in total Web impressions.
- Data starts May 18, 2026.
Clients trust reports that admit their limits.
How do you explain AI impressions to a non-technical executive?
Use a plain comparison: “An AI impression is like being quoted in a news story. It’s visibility and credibility, but not everyone who reads the story calls you.”
Then connect it to what they care about. Being cited does seem to help. Seer Interactive’s research, covering 3,100+ queries from June 2024 to September 2025, found brands cited in AI Overviews earned 35% higher organic CTR and 91% higher paid CTR than brands that weren’t cited on the same queries. That’s a correlation from one agency’s dataset, but it’s a useful reason to care about being in the answer, not just near it.
What patterns should you look for over time?
- Concentration. Is AI visibility spread across many pages or sitting on a few? Heavy concentration is a risk if one page drops.
- Page type. Are guides getting cited while service pages aren’t? That shapes the content plan.
- Device and country. Big gaps between mobile and desktop, or between markets, can point to technical or relevance issues.
- Correlation with total visibility. A study of one large health site found AI impressions tracked closely with regular search impressions by page. Its authors caution that this isn’t proof of cause, because AI impressions are already part of Web totals. Still, it suggests strong organic pages tend to be the ones AI features pull from.
How do you pair the report with GA4?
The Generative AI report tells you where you showed up. GA4 tells you what visitors did when they arrived. Together they answer the question executives actually ask: “Is this worth anything?”
A practical setup:
- Landing-page report: In GA4, open Engagement → Landing page and filter to the pages in your AI report’s top 20. Add sessions, engaged sessions and key events (form fills, calls, purchases).
- Organic segment: Compare those pages’ organic traffic this month with last month. If AI impressions rose but organic sessions fell, the AI answer may be satisfying the search without a click.
- AI assistant referrals: Build a custom channel group for sources like chatgpt.com, perplexity.ai and gemini.google.com. These aren’t Google AI Overview clicks (those count as Google organic), but they round out the AI picture.
- Self-reported source: Add a “How did you hear about us?” field to key forms. It often catches AI influence that analytics can’t.
Illustrative example (placeholders, not client data): A deck shows that a [pricing guide] had [X] AI impressions, [X] organic clicks and [X] leads last month, while a [service page] had high AI impressions but almost no clicks. The recommendation: add a clear reason to click on the service page, such as a cost calculator, local examples or a free assessment offer, and recheck next month.
What are the common mistakes?
- Leading with the total. “[X] AI impressions” means nothing without the page breakdown and what happened next.
- Double-counting. Adding AI and Web impressions.
- Claiming AI traffic from this report. It has no clicks. Use GA4 and the Performance report for traffic.
- Comparing to last year. There’s no data before May 18, 2026.
- Ignoring the opt-out setting. If a site shows zero AI impressions, check that it’s included in Search generative AI features in Search Console before diagnosing anything else.
- No actions. A slide without a next step is decoration.
What should businesses ask for in their report?
If you’re the client, ask your agency for:
- The top pages by AI impressions, with clicks and conversions next to them.
- A list of important pages that aren’t appearing, and why.
- A clear note on what the report can’t show.
- Three to five actions each month that come from the data.
Frequently asked questions
Does Search Console show AI Overview clicks?
No. The Generative AI performance report shows impressions only. Clicks from AI Overviews and AI Mode are counted inside the standard Performance report’s Web data but aren’t broken out separately.
When did the Generative AI report become available?
Google announced it in June 2026 for a limited set of sites and rolled it out to all sites worldwide on August 31, 2026. Data starts May 18, 2026.
Can I get the Generative AI report through the API?
As of September 2026, independent testing found the Search Analytics API didn’t accept a generative AI search type. Export the data from the Search Console interface instead.
Does the report separate AI Overviews and AI Mode?
No. Both features are combined into a single impressions figure in the Search report.
Why does my site show zero AI impressions?
Check that the site is included in Search generative AI features in Search Console, that key pages are indexed and snippet-eligible, and that the topics you cover actually trigger AI features. Some queries don’t.
How often should AI visibility be reported to clients?
Monthly for page-level trends and actions, with a quarterly review of patterns and strategy. Weekly reporting usually adds noise.
Want reporting your clients actually read?
We build AI visibility reporting for businesses and white-label it for agency partners, from Search Console and GA4 setup to decks that end in decisions.
Prefer to talk? Call (813) 397-3665 or email strategy@seooutsourcing.com.
