SEO & Growth

How Google Counts AI Overview Impressions in Search Console

Learn when AI Overview and AI Mode links receive impressions in Google Search Console, how position works and which reporting mistakes to avoid.

Google counts an AI Overview impression when a link to your site qualifies under the standard Search Console impression rules. A link can count even if the user does not scroll it into the browser viewport. However, if the user must expand a section to reveal the link, the link generally needs to be revealed before it receives an impression.

Every link inside the same AI Overview is assigned the position of the AI Overview block, not its individual position within the list of cited sources. These details matter because an increase in AI impressions does not necessarily mean that more people noticed a particular citation or visited the website.

What the Generative AI performance report includes

Google Search Console’s Generative AI performance report shows how often links to a website appeared within supported generative AI features on Google Search. Google currently lists two included experiences:

  • AI Overviews;
  • AI Mode.

The report groups data by page, country, device and date. Google also explains that this information comes from the Web search type in the regular Search results performance report. It is a focused view of data that is already part of Web search reporting, not an additional set of impressions to add to the main total.

Search Labs experiments are excluded while they remain under active development. The supported features and reporting methods may change as Google develops the product, so the current Generative AI performance report documentation should remain the primary reference.

What counts as an AI Overview impression?

Google describes an impression as a situation in which a user has seen, or could potentially have seen, a link to a website in Search.

The important distinction is whether the link belongs to the current set of results or whether the user must perform an additional action to reveal it.

If a link is included in the current result and the user does not need to click a control to reveal it, the link can receive an impression even if it is below the visible portion of the browser window.

This is why an impression should not be described as proof that someone read the citation. It means the link qualified as part of the served result under Google’s measurement rules.

Expandable search elements are treated differently. When a user must click a control such as “Show more” to expose additional results, links inside that hidden section generally need to be revealed before an impression is recorded.

The exact interface can change, and Google notes that its measurement heuristics may evolve. The safest interpretation is to follow the standard Search Console impression rules instead of assuming that every source associated with an AI response receives an impression.

How Google calculates position in an AI Overview

An AI Overview is treated as one search-result element. It occupies a single position, and all external links inside that element inherit the same position.

For example, if an AI Overview is recorded at position one and includes links to five different websites, those links are not reported as positions one through five. Each qualifying link receives the position of the AI Overview block.

Average position therefore does not tell you:

  • where your link appeared inside the AI Overview;
  • whether it was the first or last cited source;
  • how visually prominent it was;
  • whether the user expanded the source area;
  • whether another source attracted more attention.

Position remains useful as a broad description of where the containing search feature appeared. It should not be presented as a precise ranking among AI Overview citations.

What John Mueller clarified

The measurement question received additional attention after a discussion in the SEO community. In a Reddit response about generative AI impressions, Google Search Advocate John Mueller confirmed that the explanation based on Google’s standard impression and block-position rules was essentially correct.

He also acknowledged that traditional positions from one to ten are increasingly difficult to map onto modern search-result layouts. AI features, carousels and other grouped elements can contain several links while occupying one reported position.

That response is helpful context, but it does not replace the documentation. Public comments from Google employees can clarify how a rule applies; the maintained Help Center page remains the stronger source when the product changes.

Why the totals can be confusing

Several reporting behaviours can make AI visibility numbers look inconsistent until the aggregation rules are understood.

The Generative AI performance report is a filtered view of Web search data. Adding its impressions to the main Web search total would count the same visibility twice.

Use the report to understand the AI-related share and direction of visibility, not to create a larger combined total.

Property totals and page rows aggregate differently

Google states that if two results from the same property appear in one generative AI feature, the chart can count them as one property-level impression. When the report is grouped by page, individual pages can appear in separate rows.

This means the sum of page rows may not always match the property-level chart. That difference is expected and is not necessarily a tracking fault.

The report does not reveal every useful dimension

The documented dimensions focus on pages, countries, devices and dates. They do not provide a complete explanation of the prompts, the exact citation order or the design treatment a user saw.

An impression answers a narrow question: did a link from this property qualify for measurement inside a supported Google generative AI feature? It does not explain why the page was selected or how much influence the citation had on the user.

How businesses should report AI search visibility

AI impression data is most useful when it is treated as a trend and connected to wider business evidence.

Track the direction over time

Compare consistent periods and look for sustained changes rather than reacting to one day’s movement. Search features, reporting thresholds and preliminary data can all create short-term variation.

Annotate significant content updates, technical changes and major Google reporting anomalies. This makes later comparisons more meaningful.

Focus on the pages earning visibility

The Pages view can reveal which guides, service pages or case studies are appearing in generative AI results. Review those pages for:

  • a clear answer near the beginning;
  • accurate and attributable claims;
  • descriptive headings;
  • useful supporting detail;
  • relevant internal links;
  • an appropriate path to the next business action.

The goal is not to rewrite every paragraph for an AI system. It is to make important information easy for people and search systems to identify, verify and place in context.

Combine Search Console with analytics and enquiries

Search Console explains search visibility. Analytics can show what visitors do after reaching the website. Enquiry or sales data reveals whether that activity has commercial value.

These systems do not join AI impressions to individual conversions perfectly, so reporting should avoid false precision. A practical monthly view might include:

  1. total Generative AI impressions and their change over time;
  2. pages receiving the most AI impressions;
  3. relevant organic landing-page sessions;
  4. enquiries or other meaningful actions from organic visitors;
  5. important content or technical changes made during the period.

This produces a more honest picture than treating AI impressions as equivalent to visits.

Common reporting mistakes to avoid

Adding Generative AI impressions to Web impressions

The AI data is already included in Web search reporting. Adding both totals inflates the result.

Treating average position as citation order

All qualifying links in an AI Overview receive the block’s position. Position does not show whether your site was the first citation.

Assuming every impression was visibly read

Some links can count without being scrolled into the browser viewport. An impression represents potential visibility under Google’s rules, not confirmed attention.

Calculating an AI click-through rate from unrelated totals

Do not divide dedicated AI impressions by total Web search clicks. Those values describe different populations and would produce a misleading percentage.

Calling an impression a recommendation

A link may support one factual part of an AI response without the system recommending the company, product or service. Citation visibility and brand endorsement are different outcomes.

What this means for SEO strategy

The introduction of dedicated AI impression reporting gives website owners a useful new visibility signal. It does not change the underlying requirement to publish accurate, useful and technically accessible pages.

Google’s guidance for AI features continues to point website owners toward established search practices: allow Googlebot to crawl important content, provide indexable text, use supported structured data where it genuinely applies and make the page useful after the visitor arrives. There is no separate technical shortcut that guarantees inclusion in an AI Overview.

For most businesses, the sensible response is to strengthen the content already connected to customer questions, real services and demonstrated experience. Our SEO and digital growth service follows that evidence-led approach, while our guide to AI website development with human review explains how AI can accelerate work without replacing verification and accountability.

Primary sources

Available for selected projects

Have something
worth building?

Tell us what you are planning, what needs to improve and what success should look like.

Start a project