Google Search Console Now Tracks AI Visibility: What Businesses Can Actually Measure

The Search report currently covers impressions from AI Overviews and AI Mode. Google says the list may be updated as its generative search experiences continue to develop.

AI Overviews provide generated summaries within the main Google results page, while AI Mode gives users a more conversational way to research complicated questions, compare options and continue with follow-up queries. Although both experiences can link to external websites, they may use different systems and show different sources.

Google is initially rolling out the report to a subset of website owners so that it can test the feature and collect feedback before making it more widely available.

A business may therefore open Search Console and find that the report is missing. This does not necessarily mean the website has never appeared in AI Overviews or AI Mode. The property may not yet have access, or it may not have generated enough impressions for the report to be shown. Experimental features within Search Labs are also excluded.

The report’s primary metric is impressions. Google records an impression when a link to the website is shown to a user within a supported generative AI feature.

This gives businesses an official indication that their content has appeared within Google’s AI-generated search experiences. It can show whether visibility is increasing, declining or remaining stable. However, an impression does not confirm that a person noticed the link, visited the website or viewed the business as a preferred option.

The pages dimension shows which URLs from a website have received generative AI impressions. This can reveal that a detailed guide, comparison page, service page or research article is appearing more frequently than the homepage.

That information can help a business identify the content Google finds relevant when generating answers. A page that appears repeatedly may explain a subject clearly, provide original evidence or address a specific customer question. The report does not, however, explain why the page was selected or identify the exact section used to support the response.

Search Console can group generative AI impressions by the country where the search originated. This is useful for businesses targeting several regions or trying to understand where their content is gaining visibility.

A UK software company might discover that a large proportion of its generative AI impressions come from the United States. A consultancy may find that a service page performs strongly in Britain but has limited exposure elsewhere. This can guide decisions around regional content, international terminology and location-specific pages.

The Search report also groups impressions by desktop, mobile and tablet. This helps businesses understand which devices people are using when they encounter links to the website within AI Overviews or AI Mode.

The data does not reveal the complete customer journey, but it can provide useful context. If a high proportion of impressions comes from mobile searches, the pages being surfaced should load quickly, explain the service clearly and make the next action easy to complete. Visibility has limited commercial value when the destination page creates friction.

The report offers hourly, daily, weekly and monthly views. This allows a business to establish a baseline before publishing new content, updating an important page or implementing recommendations from an AI visibility audit.

The strongest approach is to look for sustained movement rather than expecting an immediate increase after every change. Generative search results can fluctuate as Google updates its systems, new sources are published and competitors improve their own content. Longer comparison periods provide a more reliable view of progress.

Google explains that the chart is normally aggregated at property level. If two results from the same website appear within one generative AI search feature, they may count as a single impression within the chart total. When the data is grouped by page, each URL can be analysed separately.

This can create differences between the totals shown in the chart and the table. The report is also subject to familiar Search Console limitations, including its row limit and the possibility that the newest figures are still preliminary. Businesses should therefore use it to identify patterns rather than treating every figure as a complete count of individual citations.

One of the report’s most important limitations is the absence of query data. A business cannot see the exact searches, questions or conversational prompts that caused its pages to appear.

A service page might receive hundreds of generative AI impressions without the company knowing whether they came from general research, a local enquiry, a comparison question or a high-intent request for a provider. That makes it difficult to connect the visibility directly to the customer questions the business most wants to influence.

Search Console also does not display the generated answer in which the website appeared. It cannot show whether the company was described accurately, whether an outdated service was mentioned or whether a competitor received a stronger position.

This distinction matters because visibility and representation are not the same. A business can be visible within an AI response while still being presented poorly. It may appear as a supporting source without being named, or it may be included in an answer that ultimately recommends another company.

A website appearing as a source does not automatically mean that the brand has been recommended. Google may use a page to support a definition, statistic or background statement without presenting the company as a suitable provider.

Businesses therefore need to distinguish between three different outcomes: their website being used as a source, their brand being mentioned and their company being actively recommended. The Search Console report currently measures linked appearances, not the commercial context surrounding them.

The dedicated generative AI report focuses on impressions. It does not currently provide a complete separate view of clicks, click-through rate or average position for AI Overviews and AI Mode.

Google has said that it is continuing to work with website owners and may add further metrics. For now, businesses can measure exposure more clearly than before, but they cannot use this report alone to determine how many visits, enquiries or sales came from generative AI visibility.

The new report measures appearances within generative AI features on Google Search. It does not measure how a business performs in the Gemini application, even though Gemini is also operated by Google.

It also provides no data for ChatGPT, Claude, Perplexity, Microsoft Copilot or Meta AI. These platforms can retrieve different sources, interpret questions differently and recommend different businesses. Strong visibility in Google AI Mode does not guarantee equivalent performance elsewhere.

Prompt monitoring repeatedly tests the questions that potential customers may ask AI platforms. It can show whether a company is mentioned, how it is described, which competitors appear and whether the response changes over time.

Search Console cannot provide this level of detail because it does not reveal the prompt or the full answer. A company that wants to know whether it appears when someone asks for “the best cybersecurity consultancy for a medium-sized UK business” must still test that question directly across the relevant platforms.

An impression total becomes more meaningful when it is compared with competitor performance. A business may record growing visibility while still appearing less frequently or less prominently than its closest rivals.

Competitor monitoring can reveal which organisations are repeatedly recommended, what strengths AI platforms associate with them and which sources appear to support those conclusions. None of this context is available within the new Search Console report.

Search Console identifies pages from the business’s own website that appeared within Google’s generative search features. It does not provide a complete picture of the external information shaping how the company is understood.

AI systems may also rely on directories, customer reviews, trade associations, media coverage, comparison pages and partner websites. Source analysis helps businesses understand why competitors are being trusted and where their own digital evidence may be incomplete, inconsistent or absent.

The report should be treated as the beginning of an investigation rather than the final measure of AI visibility. Businesses can identify the pages gaining impressions, compare performance across countries and devices, and monitor whether visibility changes after content or technical improvements.

Those findings should then be combined with prompt testing, recommendation tracking, competitor comparisons and source analysis. Together, these methods provide a clearer picture of whether a company is simply appearing in an AI feature or is actually being understood, trusted and recommended.

Google’s new reporting is an important development. Eligible website owners can now access dedicated first-party data showing how often their pages appear within AI Overviews and AI Mode.

However, it does not measure the complete AI discovery journey. It cannot reveal the original prompt, the generated answer, the brand’s position against competitors or its performance across other AI platforms.

AwarenessAI’s AI visibility plans combine cross-platform prompt monitoring with competitor comparisons, recommendation tracking, source analysis and prioritised opportunities. This helps businesses move beyond counting impressions and understand what AI platforms are actually saying, why competitors are appearing and where visibility can be improved.