How to Measure AI Overview and AI Mode Visibility in Google Search Console in 2026

Google Search Console now gives website owners a dedicated way to measure visibility in AI Overviews and AI Mode, but the report should not be treated like a conventional keyword-rank tracker.

Google's Generative AI performance report shows how often pages from your site appear as links in supported generative AI features, which pages appear most often, and how visibility changes by country, device and date. At the same time, AI Overview clicks, impressions and position are still part of the broader Search performance system.

That creates an important measurement rule:

Measure AI search as a funnel: visibility → cited pages → clicks → engaged sessions → conversions. Do not collapse everything into one invented "GEO score."

This guide explains what Search Console currently exposes, what it does not expose, how AI Overview position is counted, why totals can look confusing, and how a business should build a useful monthly AI-search report.

What does Google Search Console now report for AI search?

Google announced dedicated Search Generative AI performance reports on 3 June 2026.

According to Google Search Central, the Search report covers supported generative AI features including:

• AI Overviews
• AI Mode

Google says the dedicated report is being rolled out gradually, so not every Search Console property has access yet.

The report currently helps answer four practical questions:

1. How often does my site appear in Google generative AI features?
2. Which pages appear?
3. In which countries does that visibility occur?
4. Which devices are involved?

The report is useful because these questions are fundamentally different from:

"What keyword do I rank number 3 for?"

AI search can generate, expand and reformulate a user's task before displaying supporting links.

Google's own AI-search documentation says AI Mode and AI Overviews can use different models and techniques, and AI experiences may perform multiple related searches to develop an answer.

So measurement has to start with page and visibility outcomes, not only a fixed keyword list.

Where do I find the Generative AI performance report?

If your property has access, open the Generative AI performance report inside Search Console.

Google's current Search Console help documentation says the report may not appear when:

• the property has not received access yet
• the site has not received enough generative AI impressions
• the site has been excluded from Search generative AI features

Google is still rolling the report out to a subset of website owners.

Therefore:

Not seeing the report does not prove that your site has never appeared in AI Overviews or AI Mode.

It may simply mean the property has not received the dedicated reporting interface yet.

AI-feature traffic is also included in the normal Search Console Performance report under the Web search type.

What metrics are available in the dedicated AI report?

The dedicated Search Generative AI report currently centres on impressions.

Google defines an impression here as a link to your site being shown to a user in a supported generative AI feature on Search.

You can analyse those impressions by:

The report can be viewed over different date ranges and exported.

This creates a useful first dashboard:

AI impressions
      ↓
Top cited/linked pages
      ↓
Country distribution
      ↓
Device distribution
      ↓
Trend over time

But this is visibility, not business impact.

A page receiving 10,000 AI impressions and no meaningful traffic can be less valuable than a page receiving 500 impressions that produces qualified enquiries.

Does the Generative AI report show clicks?

The dedicated report launched with impression-focused visibility reporting.

Google's announcement explicitly describes impressions, pages, countries, devices and dates, and says it is considering additional metrics over time.

However, links in AI Overviews and AI Mode still participate in Google's broader Search performance accounting.

Google's AI features and your website documentation says appearances in AI Overviews and AI Mode are included in overall Search Console traffic within the Web search type.

That means you should use two complementary views:

Dedicated Generative AI report
→ AI-feature visibility

Normal Search Performance report
→ overall clicks, impressions, CTR and position

Do not assume the dedicated report is meant to replace the normal Search performance report.

It is a specialised visibility lens.

How does Google count AI Overview impressions?

Google clarified the methodology in August 2026.

Its Search Console clicks, impressions and position documentation states that an AI Overview link receives an impression when the link is actually brought into view, such as by scrolling or expanding the AI Overview.

This matters because an AI Overview can contain more content than is initially visible.

A page is not necessarily credited with an impression simply because its supporting link exists somewhere inside a collapsed area the user never sees.

So when AI impressions increase, the useful interpretation is:

Google is not merely generating answers on topics related to your site. Users are being shown links to your pages inside those experiences.

That is a stronger signal than theoretical eligibility.

How is position calculated inside an AI Overview?

This is where standard rank-tracking intuition becomes slippery.

Google's current methodology says:

An AI Overview occupies one position in the search results, and all links inside that AI Overview are assigned the same position.

Suppose the Search results page looks like:

Position 1: normal result
Position 2: AI Overview
    ├── yoursite.com/page-a
    ├── example.com/page-b
    └── another.com/page-c
Position 3: normal result

All three links inside the AI Overview are associated with the AI Overview's position.

Search Console does not tell you:

You were citation #2 inside the AI Overview card

in the same way a traditional rank tracker might claim you were the second blue link.

This means an "average AI rank" can be misleading if someone derives it without explaining Google's counting method.

Does Google give us the queries that triggered AI Overviews and AI Mode?

The dedicated Gen-AI report currently emphasises:

• pages
• countries
• devices
• dates
• impressions

It is not a full query-level "AI citation keyword" report.

That limitation is important for SEO reporting.

A third-party tool may tell you:

"You rank for 312 AI prompts."

That may be useful as its own external monitoring methodology.

But it is not the same thing as first-party Search Console data.

Google's AI optimisation guide explicitly warns site owners to be cautious with third-party tools claiming ranking success or "internal" Google metrics. Google says third parties do not have access to Google's internal ranking or AI systems.

Use third-party monitoring as supplementary evidence.

Do not present it as Google's own measurement.

Why AI visibility should be measured by page clusters

AI search often rewards pages that clearly answer a narrow task.

That makes page-level grouping particularly useful.

Imagine a software company's content produces these AI impressions:

Looking only at sitewide AI impressions gives:

10,520 impressions

That number is interesting but not very actionable.

Grouping pages by intent tells a better story:

Technical troubleshooting
→ 7,900 impressions

Commercial research
→ 1,900 impressions

Service/vendor intent
→ 720 impressions

Now you can ask:

• Is technical content earning visibility but failing to move readers toward services?
• Are commercial buyer guides underrepresented?
• Which cluster needs stronger internal linking?
• Are service pages receiving direct AI visibility?
• Which pages should be updated rather than duplicated?

This is where SEO measurement becomes business strategy.

A practical AI-search measurement funnel

A useful reporting model has five levels.

Level 1: Eligibility

Before visibility, verify basic Search eligibility.

Google's current AI-feature guidance says there are no special technical requirements for appearing in AI Overviews or AI Mode beyond being indexed and eligible to appear in Search with a snippet.

Track:

• indexed pages
• crawl/indexing issues
• canonical correctness
• snippet eligibility
• robots/meta restrictions

If a page cannot appear properly in ordinary Search, GEO tactics are not going to rescue it.

Level 2: AI visibility

From the Generative AI performance report, track:

• AI impressions
• pages receiving impressions
• country mix
• device mix
• weekly/monthly trend

This answers:

Are Google's generative experiences actually surfacing us?

Level 3: Search engagement

Use normal Search Console Performance data for:

• clicks
• impressions
• CTR
• average position
• query/page patterns

Do not try to subtract AI traffic from classic search unless Google provides a trustworthy first-party way to make that split.

Level 4: On-site engagement

Use analytics for what happens after the click.

Google recommends combining Search Console with analytics because the tools answer different questions.

Track metrics appropriate to the site, such as:

• engaged sessions
• average engagement time
• landing-page behaviour
• product/service views
• scroll depth where useful
• return visits

Level 5: Business outcomes

Ultimately track:

• qualified leads
• contact submissions
• phone calls
• booked consultations
• trials
• purchases
• revenue
• assisted conversions

This produces the complete chain:

Eligible
   ↓
Visible in AI
   ↓
Clicked from Search
   ↓
Engaged on site
   ↓
Converted

An SEO/GEO programme should improve that chain, not simply manufacture larger impression screenshots.

What should an AI visibility KPI dashboard contain?

For most businesses, keep it compact.

A useful monthly scorecard can contain:

Do not automatically add twenty metrics because Search Console can export them.

The dashboard should answer:

1. Are we becoming more visible?
2. Which content is driving that visibility?
3. Is visibility producing traffic?
4. Is traffic producing useful business outcomes?

AI impressions rising while clicks stay flat: is that bad?

Not necessarily.

Generative search can expose a brand or page without producing an immediate click.

The interpretation depends on page type.

Informational page

A user may receive enough context from the AI response that fewer clicks are required.

You should check:

• whether AI impressions are growing
• whether ordinary Search clicks changed
• whether assisted brand/search demand changed
• whether the page still drives downstream internal navigation

Commercial page

If a buyer-intent page receives large AI visibility but almost no traffic or leads, inspect whether:

• the title/description promise a clear reason to click
• the page contains detail beyond the AI summary
• the page answers the next decision, not just the definition
• the CTA matches user intent
• the content is being cited for informational facts but not commercial comparison

The goal is not to force every AI impression into a click.

The goal is to earn clicks where a click creates additional value.

Why CTR can become harder to interpret

CTR is still useful, but context matters.

An AI Overview may appear above or between normal results and can expose several external links.

A rise in AI impressions can change the denominator of Search visibility without producing proportional click growth.

Therefore review:

CTR
+
page intent
+
AI impression trend
+
conversion trend

rather than declaring:

"CTR fell, therefore SEO performance is worse."

A lower CTR with dramatically higher qualified conversions can still be a business improvement.

Likewise, higher AI impressions with falling conversions may indicate the site is appearing for broader informational contexts that do not match the commercial objective.

Why Search Console chart and table totals can differ

Google's Gen-AI Search report uses different aggregation depending on the view.

According to Google's help documentation:

• the chart is aggregated by property
• page tables are aggregated by page

If two URLs from the same site appear in one generative AI result, the property-level chart may count that as one site impression while page-level rows can attribute impressions to individual pages.

That means:

sum(page rows)

does not always equal:

sitewide chart total

This is normal Search Console aggregation behaviour, not necessarily a reporting bug.

Always note the aggregation level before copying figures into a client report.

What does "canonical URL" mean in the AI report?

Page-level performance is generally attributed to canonical URLs.

That matters when a site contains:

• duplicate product URLs
• UTM parameters
• alternate mobile URLs
• print views
• tracking parameters
• duplicate category paths

You may expect an alternate URL to appear in the report but see the canonical page instead.

This is another reason page-level AI reporting works best when the site's canonical strategy is already clean.

Poor technical SEO makes AI-performance interpretation harder.

What date and timezone does the report use?

Google's Search Console Gen-AI report uses Pacific Time for dates, consistent with broader Search Console reporting conventions.

This matters for teams comparing Search Console to:

• GA4
• CRM records
• server logs
• sales dashboards
• local-day reporting

If your business operates in London, Karachi, Dubai or Sydney, a "day" in Search Console may not line up exactly with the same calendar day in your internal system.

For daily analysis, note the timezone.

For monthly trend reporting, the difference is usually less material.

What if my site does not have the Gen-AI report yet?

Use the data you do have.

1. Normal Search Console Performance report

Google says AI Overviews and AI Mode are included in overall Web Search performance.

2. Landing-page analysis

Track whether content designed for AI-friendly retrieval is gaining impressions, clicks and conversions overall.

3. Google Analytics

Measure what organic visitors do after arrival.

4. Manual/third-party monitoring

Use carefully for research and prompt discovery, but label it correctly.

For example:

"Observed in our monitoring sample"

is different from:

"Google reports 1,200 AI citations."

Do not blur the two.

Should we optimise specifically for the Generative AI report?

Do not optimise for the dashboard.

Optimise for users and Search.

Google's current AI optimisation guidance says the same foundational SEO practices remain relevant for generative AI features.

That includes:

• useful original content
• clear textual information
• crawlability
• internal linking
• good page experience
• accurate structured data where appropriate
• useful images/video for relevant queries
• up-to-date merchant/local information where applicable

There is no special "AI Overview schema" that guarantees inclusion.

There is no hidden __INLINE_CODE_0__ meta tag.

The reporting layer tells you which pages Google's systems are already finding useful in these experiences.

Use that feedback to improve content strategy.

A practical monthly GEO review

Step 1: export AI impressions

From the Gen-AI report, export:

• page
• AI impressions
• country
• device
• date where useful

Step 2: classify pages by intent

For example:

Technical troubleshooting
Commercial comparison
Cost/buyer guide
Service page
Local page
Research/explainer

Step 3: identify winners

Look for:

• fast-growing AI impressions
• pages appearing across multiple countries
• service/commercial pages gaining visibility
• clusters with several cited pages

Step 4: identify gaps

Look for:

• high-value services with no AI-visible supporting content
• content clusters generating visibility but no conversion path
• important pages with declining impressions
• repeated articles competing for the same intent

Step 5: connect Search Console with conversions

For AI-visible landing pages, review:

• organic sessions
• engagement
• CTA interactions
• contact submissions
• assisted conversions

Step 6: update instead of cloning

If a page is gaining AI visibility but missing key follow-up questions, improve the existing page.

Do not create five near-identical pages targeting slightly different phrasings.

Step 7: record the hypothesis

Example:

Observation:
Firebase troubleshooting pages receive high AI visibility.

Hypothesis:
Google values Softotic's diagnosis-first technical content.

Action:
Publish another genuinely distinct production troubleshooting article.

Success measure:
New article gains AI impressions and qualified technical-service visits
without reducing visibility of existing Firebase pages.

This turns GEO into an iterative content programme instead of a keyword costume party.

What not to report to stakeholders

"We rank #1 in AI"

Unless the methodology is explicitly defined, this is usually too vague.

Which feature? Which location? Which prompt? Which model? Which session? Which measurement source?

"Our AI visibility score is 82"

If it is a proprietary score, label it as proprietary.

Do not imply Google created it.

"AI Overviews generated exactly X clicks"

Only say this when the first-party reporting actually gives you a defensible way to isolate those clicks.

"Search Console's new report caused our impression spike"

Google's Gen-AI data was already included in broader performance totals; introducing a dedicated report did not simply inject an entirely new class of impressions into historical Search performance.

"More AI impressions always mean better SEO"

Visibility without relevant traffic or business outcomes can still be low-value.

How should agencies report GEO performance?

A credible client report should separate facts, observations and hypotheses.

For example:

First-party fact

Search Console reported 18,200 generative AI impressions this month.

First-party breakdown

62% of page-level AI visibility came from technical troubleshooting content.

Analytics fact

Organic visitors landing on those pages generated 17 qualified enquiries.

Interpretation

Technical authority is currently earning more AI visibility than bottom-of-funnel service content.

Recommendation

Build internal links and decision-stage articles around the strongest technical clusters rather than publishing duplicate troubleshooting pages.

That reporting style is much more useful than:

"GEO increased by 34%."

The metric should teach you what to do next.

FAQs

Does Google Search Console show AI Overview impressions?

Yes. Google's Generative AI performance report includes impressions from supported generative AI features in Search, currently including AI Overviews and AI Mode.

Does Search Console show AI Mode separately from AI Overviews?

The dedicated report currently groups supported Search generative AI features into the Gen-AI reporting experience rather than acting as a conventional per-keyword rank tracker for each AI product. Google may add metrics or capabilities as the report evolves.

How is an AI Overview link's position counted?

Google says the AI Overview occupies one position in Search results and all links inside the AI Overview receive that same position.

Does an AI Overview link get an impression if it is hidden?

Standard impression rules apply. Google says the link must be scrolled or expanded into view to count as an impression.

Why can AI report page totals differ from the chart?

The chart can aggregate by property while a page table aggregates by individual page. Multiple pages from one property in the same generative result can therefore create differences between chart and table totals.

Is there a special GEO ranking factor or schema for AI Overviews?

Google says there are no additional technical requirements or special optimisations required to appear in AI Overviews or AI Mode beyond normal Search eligibility and SEO best practices.

Conclusion

Google's 2026 Generative AI performance reporting finally gives website owners a first-party way to see whether their pages are surfacing inside AI Overviews and AI Mode.

But the most useful metric is not a new rank number.

Measure the whole path:

Search eligibility
        ↓
AI impressions
        ↓
Pages and content clusters
        ↓
Search clicks
        ↓
On-site engagement
        ↓
Qualified conversions

Then use the data to decide what to improve, update, connect or stop producing.

The businesses that get the most value from GEO measurement will not be the ones with the fanciest proprietary score.

They will be the ones that can answer:

Which pages are earning AI visibility, why are they useful, and does that visibility contribute to real business outcomes?

If you need to connect technical SEO, content strategy and AI-search measurement into one system, Softotic's SEO and growth optimisation service can help, while web application development can support analytics, reporting and custom measurement workflows.

Sources and references

• Google Search Central: Introducing Search Generative AI performance reports in Search Console
• Search Console Help: Generative AI performance report (Search)
• Search Console Help: What are impressions, position and clicks?
• Google Search Central: AI features and your website
• Google Search Central: Optimising for generative AI features on Google Search
• Google Search Central: Using Search Console and Google Analytics data for SEO