Measure AI Search Visibility in Google Search Console

You can now measure Google AI-search visibility directly in Search Console’s generative AI performance report. The report shows impressions from AI Overviews and AI Mode, then lets you compare the pages, countries, and devices generating those appearances. It does not turn visibility into a single ranking score, so the useful workflow is to pair its impression data with normal Search performance and on-site outcomes.

For a small business, the practical question is not simply whether a page appeared in an AI response. It is whether the right pages are being surfaced for commercially relevant topics and whether search visitors continue to a useful action such as reading a guide, viewing a service, requesting a quote, or buying a product.

KEY TAKEAWAYS

What you need to know

Google’s generative AI report covers impressions from AI Overviews and AI Mode.
Use page, country, device, and date views to find where AI-search visibility is changing.
An impression is exposure, not proof of a click, lead, sale, or ranking improvement.
Measure business value by combining Search Console visibility with landing-page engagement and conversions.
Compare stable reporting periods and annotate major content or technical changes before drawing conclusions.

What does Google’s generative AI report measure?

The report measures how often links to your site receive impressions in supported generative AI features on Google Search. Google currently lists AI Overviews and AI Mode as included features. The default chart shows change over time, while the available dimensions help identify which pages appeared and where those impressions originated.

Google says the report was rolled out to websites worldwide on August 31, 2026. A property may not see the report when it has not accumulated enough generative AI impressions. Absence of the report is therefore not proof of a technical fault or an exclusion.

This is a visibility report. Treat it as evidence that Google displayed a link to your site within a supported AI experience—not as a complete account of the user journey or a standalone measure of SEO success.

Continue reading: Google Search Console: generative AI performance report ↗ · SEO content knowledge hub

How AI-search impressions are counted

Google applies its standard impression rules to AI Mode. For AI Overviews, the overview occupies one position in the results and links inside it receive that same position. In AI Mode, a follow-up question is treated as a new query, so impressions, positions, and clicks in the new response are counted against that follow-up query.

That methodology makes raw position difficult to interpret without context. One page can appear as a supporting link inside an overview, while another may be surfaced during a deeper follow-up. The number alone does not explain the wording, prominence, user intent, or competitive set behind the appearance.

What each available signal can and cannot tell you
Signal	Useful for	Do not assume
Impressions	Tracking AI-search exposure over time	That a user clicked or read the page
Pages	Finding URLs surfaced in AI experiences	That every impression served the same intent
Countries	Seeing geographic distribution	That location alone explains performance
Devices	Comparing mobile and desktop exposure	That device mix proves conversion quality
Position	Understanding counted placement under Google’s rules	A directly comparable traditional blue-link rank

Continue reading: Google: how AI Mode and AI Overview metrics are counted ↗ · Search intent definition

A practical measurement workflow

Start with a clear question and a fixed comparison period. For example: did the pages updated during October gain generative AI impressions compared with the previous four weeks? Record the dates of major content, canonical, indexing, template, or internal-link changes so a later increase is not credited to the wrong action.

1. Establish a baseline

Export or record total impressions for a period long enough to smooth out daily noise. Use the same property type and filters every time. If the report is new for the site, label the first period as a baseline rather than declaring a trend from a few days of data.

2. Review landing pages

Sort or compare by page to identify which URLs earn visibility. Group them by role: glossary definition, comparison, guide, product, service, location, or news analysis. A cluster view is more actionable than a list of unrelated winners because it shows whether Google is finding the parts of the site intended to answer a topic.

3. Segment by country and device

Use country and device dimensions to test a specific business question, such as whether a localized page is visible in its target market or whether a mobile-heavy audience sees the same pages as desktop users. Avoid inventing explanations from a segment without checking content, demand, and site experience.

4. Compare with normal Search performance

Review the same pages in the standard Search performance data. Look for changes in clicks, impressions, queries, and click-through rate, while remembering that the generative AI report is a separate view with its own scope. The goal is to see whether AI exposure is part of wider organic discovery or concentrated on a small set of pages.

5. Connect visibility to outcomes

In analytics or your own first-party reporting, examine engaged sessions, useful next-page visits, sign-ups, enquiries, purchases, or other outcomes for the surfaced landing pages. Do not present correlation as causation when Search Console cannot identify every visit as coming from a particular AI feature.

Continue reading: Google AI Overviews optimization guide

Which pages should you prioritize?

Prioritize pages that already answer a distinct reader need and connect naturally to a next step. A glossary page can define an entity, a comparison can support a choice, a guide can explain a process, and a commercial page can present a relevant offer. Publishing many near-duplicate pages for possible fan-out questions can dilute the site and conflict with Google’s guidance against scaled, low-value content.

A useful review asks whether the page adds information an AI response cannot fully replace: first-hand experience, a tested workflow, an original calculation, clear constraints, current availability, proprietary data, a decision framework, or a service the reader can actually use.

High-value pages with accurate, current information.
Pages receiving AI impressions but weak internal navigation.
Strong pages missing descriptive images, evidence, or clear authorship.
Pages whose intent overlaps another URL and needs consolidation.
Pages visible in a target country but not yet localized for that audience.

Continue reading: SEO vs SEM comparison · Google’s guidance for generative AI search ↗

How to interpret gains and declines

A rise in impressions can mean more eligible pages, stronger relevance, changing demand, wider feature availability, or a shift in how Google presents sources. A decline can reflect the reverse, but it can also come from seasonality, reporting thresholds, changed query behavior, or a smaller set of AI responses. Investigate before assigning a cause.

Compare periods of similar length, review affected pages, and check whether the change is broad or limited to a country, device, or content group. Then inspect the page itself: indexability, canonical, freshness, accuracy, main content, images, internal links, structured data, and page experience. Google states that pages must be indexed and eligible to appear with a snippet to be eligible as supporting links in AI Overviews or AI Mode.

Continue reading: Google: AI features and your website ↗

What not to report as success

Avoid turning a small impression increase into a claim that revenue, authority, or rankings improved. Do not combine unrelated dates until the graph looks favorable. Do not call a manual screenshot a repeatable measurement method, because AI results can vary with the query, follow-up, location, device, and product experience.

Also avoid creating an invented ‘AI visibility score’ unless its formula, inputs, limitations, and sampling method are disclosed. Search Console provides first-party impression data; any blended score layered on top is a separate model, not a Google metric.

Impressions without a stable comparison period.
One manually checked query presented as market coverage.
Position treated as a directly comparable classic organic ranking.
Traffic or revenue attributed to AI features without supporting evidence.
More indexed pages presented as proof of better AI-search visibility.

A monthly reporting template for small businesses

A short monthly review is usually more useful than a large dashboard nobody acts on. Keep the report tied to decisions: which pages deserve an update, which topic has earned deeper coverage, which country needs localization, and which high-impression page needs a clearer commercial next step.

Suggested monthly AI-search visibility review
Report item	Question to answer	Action
Total AI impressions	Is exposure rising, falling, or stable?	Investigate material changes
Top gaining pages	Which page roles and topics grew?	Strengthen evidence and internal links
Top declining pages	Is the loss technical, editorial, or demand-led?	Inspect before rewriting
Country and device mix	Does visibility match the target audience?	Improve localization or experience
Landing-page outcomes	Do surfaced pages lead to useful actions?	Clarify next steps and measure conversions
Changes shipped	What was published or fixed this month?	Annotate the timeline for later comparison

Continue reading: 90-day local SEO plan · Google’s October 2026 AI Search link updates ↗

The decision to make after each review

End every review with one documented decision. You might improve a high-impression guide, consolidate overlapping pages, add a missing comparison, strengthen a page’s next step, or leave a stable page alone. The report is valuable when it changes priorities—not when it simply produces another chart.

For most small businesses, the best next action is to improve a small number of pages with proven visibility and clear commercial relevance. Preserve their distinct intent, make the information more useful and current, and measure the next period with the same filters.

COMMON QUESTIONS

Frequently asked questions

Does Search Console show AI Overview and AI Mode data separately?

Google’s generative AI performance report currently includes impressions from both AI Overviews and AI Mode. Use the dimensions available in the report, but do not assume every impression can be assigned to one feature unless Google explicitly provides that breakdown in your interface.

Why can’t I see the generative AI report?

Google says the report may not appear when a site has not received enough impressions in generative AI features. Also confirm that the property is eligible and included in Search generative AI features.

Are AI-search impressions the same as clicks?

No. An impression means a link to your site was shown under Google’s counting rules. A click is a separate action, and a conversion is another step beyond that.

Should a business create pages for every AI fan-out query?

No. Google advises against publishing many low-value pages to target query variations. Build distinct pages only when each one serves a genuine reader need and adds useful, non-commodity information.