# Google Search Console Generative AI Features (Beta): How to Track AI Overviews and AI Mode Visibility in 2026 Learn how Google's Generative AI Features report tracks AI Overviews, AI Mode, impressions, pages, and visibility, plus practical SEO tips for 2026. Canonical URL: https://aiidelist.com/blog/google-search-console-generative-ai-features Language: en Published: 2026-10-17 Updated: 2026-10-17 Google Search Console has introduced a dedicated **Generative AI Features (Beta)** performance report, giving website owners a clearer view of how their content appears in Google's AI-powered search experiences. For years, publishers could measure traditional search rankings, impressions, clicks, and click-through rates. However, understanding whether a website appeared inside AI-generated answers was considerably more difficult. The new report addresses part of that measurement gap. It reveals which pages receive exposure in Google AI Overviews and AI Mode, where those impressions originate, and how visibility changes over time. Importantly, it does not yet provide a complete picture of AI-driven traffic or conversions. This guide explains what the report measures, how to interpret its limitations, and how website owners can turn the available data into actionable SEO decisions. ## Key Takeaways - **Google launched Generative AI performance reporting on June 3, 2026**, with worldwide availability announced on August 31, 2026. - **The report tracks organic impressions from AI Overviews and AI Mode**, with separate reporting for generative AI experiences in Google Discover. - **Website owners can analyze impressions by page, country, device, and date.** Device breakdowns are available in the Search report. - **AI-specific clicks, CTR, average position, and search queries are not provided** in the dedicated Generative AI report. - **AI impressions are already included in existing Search Console performance data.** Adding the new report's impressions to regular Search impressions would double-count overlapping exposure. - **A September 24, 2026 update introduced multimodal search reporting**, covering image-based searches through experiences such as Google Lens and Circle to Search. - **The most effective optimization strategy remains strong technical SEO and genuinely useful original content**, rather than specialized AI markup or unsupported GEO tricks. ## What Is Google Search Console Generative AI Features (Beta)? Generative AI Features is a performance reporting section inside [Google Search Console](https://search.google.com/search-console/about). It measures how frequently links to a verified website appear in Google's supported generative AI search experiences. The report currently covers two major Google Search features: **Google AI Overviews** AI Overviews are generated summaries that appear within Google Search results for eligible queries. They combine information from relevant sources and may present links to supporting webpages. For example, a user searching for the best daily geography guessing games might receive an AI-generated comparison containing links to several game directories or individual games. If a website's URL appears in the supported AI experience and meets Google's impression-counting requirements, that appearance can contribute to its Generative AI performance data. **Google AI Mode** AI Mode provides a more conversational search experience. Users can ask complex questions, explore comparisons, and submit follow-up questions. Google may retrieve information from multiple sources and display supporting links within the generated responses. Website owners can now measure aggregated visibility from these experiences without relying exclusively on manual searches or third-party AI monitoring services. The report is primarily a **visibility measurement tool**, not a separate ranking system, AI content generator, or traffic attribution platform. Google also provides a separate [Generative AI performance report for Discover](https://support.google.com/webmasters/answer/16983858), where content may appear in AI-powered discovery experiences outside conventional keyword searches. ## When Did Google Launch the Generative AI Report? The rollout progressed through several important milestones in 2026. | Date | Development | Why It Matters | |---|---|---| | June 3, 2026 | Google announced Generative AI performance reports | Publishers gained a dedicated view of AI search impressions | | August 31, 2026 | Worldwide rollout completed | The reporting capability became globally available, subject to property-level data eligibility | | September 24, 2026 | Multimodal search reporting introduced | Image-based search visibility became easier to distinguish from text-based search | | October 15, 2026 | Google announced more prominent source links in AI responses | Publishers gained additional potential opportunities for users to discover and visit original sources | Google's [original Search Central announcement](https://developers.google.com/search/blog/2026/06/gen-ai-performance-reports) confirms the global rollout and the report's supported dimensions. Although the interface may still display a Beta label, the feature is no longer restricted to the original testing group. However, global availability does not guarantee that every website will have enough qualifying impressions to display a useful report. ## What Data Does the Generative AI Report Show? The report focuses on measurable appearances of website links within supported AI features. ### 1. AI Impressions **Impressions are the primary metric.** They represent instances in which links to a website were shown to users in Google's supported generative AI search experiences. For example, if a website receives 2,000 Generative AI impressions during a seven-day period, its links were displayed in qualifying AI search results according to Google's reporting rules. This does not establish that 2,000 different people visited the website. An impression is not equivalent to a click, unique visitor, conversion, or successful recommendation. ### 2. Pages The Pages dimension identifies which website URLs appeared in supported generative AI results. This helps publishers answer questions such as: - Which articles appear most frequently in AI-generated answers? - Are product comparisons receiving more exposure than individual product descriptions? - Do tutorials attract AI visibility even when their traditional search rankings are modest? - Which content categories should receive further editorial investment? Google generally assigns performance data to the canonical URL rather than duplicate versions of the same content. The report also accounts for the final linked URL after redirects, making canonicalization an important consideration when investigating unexpected page-level results. ### 3. Countries The Countries dimension identifies where AI impressions originate. For international websites, this can reveal whether content is reaching the intended geographic audience. A technical article might receive most of its AI impressions from the United States, while a multilingual travel guide could generate substantial visibility in several different markets. Country-level data can help guide localization, regional content improvements, and market prioritization. ### 4. Devices For Google Search, website owners can compare impressions across desktop, mobile, and tablet devices. Differences may reflect search behavior, feature availability, interface presentation, and audience composition. However, higher impressions on one device category do not automatically indicate stronger rankings or higher engagement. The separate Discover Generative AI report does not provide the same device dimension. ### 5. Dates and Trends Date-based reporting helps identify changes in AI search visibility. Publishers can investigate whether exposure increased after publishing a comprehensive guide, updating an outdated comparison, improving indexing, or expanding coverage of a topic. Google supports multiple time granularities, allowing short-term changes and broader trends to be examined. Search Console reports use **Pacific Time**, which is important when comparing the data with analytics platforms configured to different time zones. Recent data may also be preliminary and subject to revision. ## What the Report Does Not Show The most important limitation is that Generative AI reporting does not currently offer the same performance metrics as the standard Search results report. | Metric | Generative AI Search Report | Standard Search Results Report | |---|---|---| | Impressions | Yes | Yes | | Pages | Yes | Yes | | Countries | Yes | Yes | | Devices | Yes | Yes | | Date trends | Yes | Yes | | Clicks | No AI-specific breakdown | Yes, including blended Web search traffic | | Click-through rate | No | Yes | | Average position | No | Yes | | Search queries | No | Yes, without dedicated AI attribution | | Separate AI Overview and AI Mode totals | No | No dedicated separation | | AI-generated answer text | No | No | | Exact citation placement | No | No | This distinction matters because a website can gain AI visibility without receiving a proportional increase in organic traffic. A page could appear repeatedly as a supporting source, while users obtain sufficient information from the generated answer and never open the page. Conversely, an AI response could send highly motivated users to a specialized tool, research report, or comparison page. The current report cannot directly distinguish those two outcomes. **AI impressions should therefore be treated as an early visibility indicator rather than a direct measure of business performance.** ## How to Find Generative AI Features in Google Search Console Website owners with a verified Search Console property can access the report through the existing interface. 1. Open [Google Search Console](https://search.google.com/search-console). 2. Select the relevant website property. 3. Locate the **Performance** section in the left navigation. 4. Open **Generative AI features**, which may display a Beta label. 5. Select the desired date range. 6. Review the impressions chart and examine the Pages, Countries, Devices, and Dates dimensions. 7. Use the available filters to refine the dataset. 8. Select **Export** to download the displayed chart and table data for further analysis. Google also provides [official documentation for the Search report](https://support.google.com/webmasters/answer/16984139). For a meaningful first analysis, compare the previous 28 days with the preceding 28-day period, using complete days rather than relying on potentially incomplete recent data. For newer websites or properties with limited exposure, longer observation periods may produce more useful signals. ## How Google Counts AI Search Impressions Understanding aggregation is essential because Search Console's chart totals and page-level tables do not always match. ### Property-Level Aggregation The main Search report chart generally aggregates impressions at the property level. Suppose Google displays two links from the same website in a single qualifying generative AI search result. At the property level, Google counts that exposure as **one impression**, not two. ### Page-Level Aggregation When viewing results grouped by page, different qualifying URLs can be counted individually. Consider the following hypothetical AI response: - `example.com/guides/ai-image-models` - `example.com/tools/image-comparison` If both URLs appear in the same qualifying result, the site-level chart may record one impression while the page table attributes an impression to each URL. This does not necessarily indicate a reporting error. It reflects different aggregation methods. ### Search and Discover Have Different Counting Rules Discover's Generative AI report uses page-based aggregation. Multiple qualifying results from the same property in a Discover list can each count separately. Discover impressions also depend on content being scrolled into view, and repeated viewing of the same result within a session does not automatically generate additional impressions. Because the reporting systems measure different experiences and apply different counting conventions, Search and Discover figures should be analyzed separately before drawing cross-platform conclusions. ## Are Generative AI Impressions Already Included in Regular GSC Data? **Yes. Google states that Generative AI Search impressions are already included in the standard Search performance report under Web search.** The dedicated report provides an additional view of information already represented in broader Search reporting. Consequently, publishers should not add the Generative AI impression total to standard Search impressions as if they represented two independent traffic sources. For example, consider a hypothetical website with: | Metric | Monthly Value | |---|---:| | Total Web search impressions | 100,000 | | Generative AI Search impressions | 6,000 | | Total Web search clicks | 3,500 | A useful descriptive calculation is: ```text AI impression share = Generative AI impressions / Total Web impressions × 100 6,000 / 100,000 × 100 = 6% ``` This suggests that reported AI appearances represent approximately 6% of the site's overall Web search impressions, subject to the reports' aggregation rules and comparable date ranges. It does **not** mean that 6% of visitors arrived through AI search. Nor does it establish that 6% of revenue was generated through AI-powered results. The distinction between visibility and traffic is fundamental to interpreting the new reporting system correctly. ## September 2026 Update: Multimodal Search Reporting On September 24, 2026, Google announced an additional reporting capability for multimodal search. The update introduced a search-type filter that distinguishes conventional text-based Web searches from supported image-based search experiences. The available categories include: **Web: text-based** Searches originating from text queries entered into the standard Google search interface. **Web: multimodal** Searches in which users provide images as part of the search interaction, including supported experiences such as: - Google Lens - Circle to Search on Android - Images uploaded to Google Search - Chrome's image-search functionality This distinction is relevant to websites that publish visually distinctive material, including image-editing tools, design tutorials, product catalogs, visual puzzle games, and travel guides. For example, a guide explaining how to identify an unfamiliar flower could gain visibility when users submit photographs rather than typing the flower's name. Publishers should evaluate image quality, descriptive context, accessibility, and the usefulness of surrounding content. The new filter does not mean every image impression originates in an AI-generated answer. The search type and the reporting surface are separate dimensions that must be interpreted together. Google describes the update in its [official multimodal reporting announcement](https://developers.google.com/search/blog/2026/09/web-multimodal-in-sc). ## How to Analyze Generative AI Data for SEO Opportunities The greatest value of the new report comes from connecting visibility trends with content strategy. A practical analysis can be organized into five stages. ### Step 1: Identify Your Most Visible Pages Open the Pages dimension and sort URLs by Generative AI impressions. Group them by content type, such as: - Product reviews and comparisons - Technical tutorials - Educational explanations - Tools and calculators - Game descriptions and guides - Original research and datasets Look for patterns rather than focusing only on a site's highest-volume URL. If comparison pages consistently receive more AI impressions than generic product listings, this may indicate that users are asking questions requiring evaluation, context, and decision support. That interpretation is a hypothesis, not proof of Google's internal ranking preferences. ### Step 2: Compare AI Exposure With Conventional Search Performance For each high-visibility page, examine its performance in the standard Search results report. Useful questions include: - Has overall organic traffic increased alongside AI exposure? - Did impressions grow while clicks remained relatively stable? - Does the page rank for informational searches but receive limited commercial engagement? - Are relevant pages attracting users from the intended markets? Because AI-specific clicks are not isolated, these comparisons are directional rather than definitive attribution. ### Step 3: Find Content With Growth Potential Pages with rising AI impressions but limited conventional clicks deserve investigation. Possible explanations include strong informational relevance, limited click incentives, competition from more established domains, or changes in Google's result presentation. Useful improvements may include adding original benchmarks, interactive tools, updated screenshots, downloadable resources, or comparisons that genuinely require users to visit the website. The goal should not be to withhold basic information artificially. Instead, publishers should deliver additional value that cannot be fully reproduced in a short search summary. ### Step 4: Compare Countries and Devices Regional and device breakdowns can identify differences in visibility. For example, substantial impressions in a country where a website has no localized content may reveal an opportunity to improve language support and regional relevance. Likewise, a website primarily accessed on mobile devices should prioritize responsive layouts, fast loading, and usable interfaces. These are established SEO and usability principles, not unique requirements of AI search. ### Step 5: Track Results Over Several Weeks AI visibility can fluctuate as search demand, indexing, model behavior, and Google's interface evolve. A weekly measurement process is more reliable than making major content decisions based on one day's impressions. The following reporting template can help: | KPI | Current Period | Previous Period | Recommended Review | |---|---|---|---| | AI Search impressions | Record total | Record total | Calculate growth | | Pages receiving AI impressions | Count URLs | Count URLs | Identify coverage changes | | Top 10 AI-visible pages | Record URLs | Record URLs | Compare page movement | | Leading countries | Record shares | Record shares | Assess target markets | | Standard Web clicks | Record total | Record total | Check broader traffic trends | | GA4 organic conversions | Record total | Record total | Evaluate business outcomes | When using GA4, remember that its organic search measurements cannot independently identify every visit originating from a specific AI Overview or AI Mode citation. Traffic changes and AI exposure trends can be compared, but correlation should not be presented as confirmed causation. ## Why Isn't the Generative AI Report Showing Up? Some verified websites may not see the report or may have insufficient data to analyze. Google identifies inadequate qualifying impressions as a common explanation. Other conditions worth investigating include: **1. Limited AI visibility** The website may not yet have generated enough impressions in supported AI search features. **2. Generative AI exclusion settings** The property may be excluded from supported Search generative AI features. **3. Property selection** The selected Search Console property might differ from the canonical URLs receiving impressions. **4. Indexing or crawlability issues** Pages that are not indexed or are ineligible to appear with snippets in Google Search cannot meet the standard eligibility requirements for supported AI Search appearances. **5. Unsupported experimental experiences** The dedicated report excludes data from experiments in Google Search Labs. **6. Data collection and reporting limitations** Newly collected data may still be preliminary, while low-volume dimensions may not provide complete granular breakdowns. Not seeing a report should not automatically be interpreted as a technical penalty or proof that Google dislikes a website's content. ## How to Improve Visibility in Google AI Overviews and AI Mode Google's [official guidance for generative AI search optimization](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide) emphasizes that traditional SEO fundamentals remain important. The company describes AI-powered search as an extension of its existing retrieval and ranking systems rather than an entirely separate optimization ecosystem. Two technical concepts help explain this relationship. ### Retrieval-Augmented Generation Retrieval-augmented generation, or RAG, uses retrieved information to help generate grounded responses. Google's AI Search experiences can rely on relevant webpages from the Search index to provide supporting information and links. This makes crawlability, indexing eligibility, topical relevance, and trustworthy content important. ### Query Fan-Out Query fan-out allows an AI system to investigate related subquestions when answering a broader request. For example, someone asking which AI image generator is best for a low-memory GPU might trigger research involving model requirements, benchmark comparisons, pricing, and installation procedures. A detailed resource covering these decision points may be useful across several related information needs. However, creating hundreds of near-identical pages targeting every possible question variation is not a recommended strategy. ### Practical Optimization Checklist - **Publish original information:** Include verified specifications, useful comparisons, original research, or genuine experience-based insights. - **Maintain indexability:** Ensure important pages can be crawled, indexed, and displayed with eligible snippets. - **Use descriptive headings:** Organize pages around meaningful subjects and user needs. - **Make important content accessible:** Avoid relying entirely on client-side rendering that could prevent content from being processed reliably. - **Handle canonical URLs correctly:** Reduce unnecessary duplication and conflicting canonical signals. - **Use high-quality supporting media:** Add useful diagrams, screenshots, images, and videos where appropriate. - **Update time-sensitive facts:** Keep pricing, compatibility, specifications, and product availability current. - **Strengthen internal linking:** Help visitors and search engines discover related resources. - **Deliver usable page experiences:** Improve mobile usability, navigation, and loading performance. - **Evaluate outcomes:** Measure traffic, engagement, conversions, and recurring AI visibility rather than impressions alone. There is no guaranteed technique for earning a citation in a specific AI-generated response. ### Do Websites Need Special AI Schema or an llms.txt File? No special schema.org markup is required solely to appear in Google's generative AI search results. Google also states that it does not use `llms.txt` as a special discovery or ranking mechanism for Google Search. Structured data remains useful when accurately applied to supported Search features, but adding unnecessary markup does not guarantee AI inclusion. Similarly, forcing every article into tiny AI-oriented content chunks or manufacturing artificial online mentions is not a reliable optimization strategy. Publishers should prioritize useful content and technical accessibility over unverified AI SEO shortcuts. ## Can Website Owners Disable Google Generative AI Features? Yes. Google introduced a dedicated Search generative AI inclusion control in Search Console in 2026. It is available under: ```text Google Search Console → Settings → Search generative AI ``` Website owners can choose to include or exclude their content from supported generative AI search features. **Include** The site's eligible links and content may appear in supported features, including AI Overviews, AI Mode, and generative AI experiences in Discover. This is the default setting for top-level properties unless an applicable inherited configuration changes the result. **Exclude** The site's links and content become ineligible for inclusion and grounding in the supported AI features. Google states that excluded sites will not receive impressions or traffic from those features. Property settings can also be inherited from parent properties, so administrators managing subdomains or URL-prefix properties should review the effective configuration carefully. Importantly, the control does **not** serve as a ranking signal for other Google Search results. It is also distinct from controls affecting model training. Google identifies `Google-Extended` as a separate mechanism for limiting certain AI model training uses, while `noindex` affects inclusion in Google Search more broadly. Google notes that inclusion-control changes generally take effect within days, although caching and system propagation may introduce additional delays. Additional details are available in the [Search generative AI control documentation](https://support.google.com/webmasters/answer/16908024). For publishers dependent on organic discovery, excluding the entire website should be a deliberate business decision rather than a routine SEO adjustment. ## October 2026 Update: Google Makes AI Source Links More Prominent On October 15, 2026, Google announced additional improvements to how source links appear in AI Overviews and AI Mode. The changes include: - More inline links placed directly within generated responses. - Greater prominence for supporting links appearing earlier in answers. - Testing of dedicated sections highlighting relevant official websites. - Expanded article carousels for timely information in more locations and languages. - More visually engaging links with thumbnails in certain AI responses. - Testing of clickable creator and publisher profiles in video-related AI experiences. These updates are designed to encourage users to explore original sources rather than relying exclusively on AI-generated summaries. For website owners, the changes reinforce an important distinction: **being visible in an AI answer and earning a click from that answer are different outcomes.** Google's reporting improvements make exposure easier to monitor, while changes to source presentation may affect how frequently users visit linked websites. However, the dedicated Generative AI report still does not provide AI-specific click attribution. The full announcement is available in Google's [October 15 AI Search update](https://blog.google/products-and-platforms/products/search/updates-ai-mode-overviews-explore-web/). ## Common Reporting Mistakes to Avoid Several mistakes can lead to misleading SEO conclusions. ### Mistake 1: Treating Impressions as Website Visits A displayed link does not establish a completed visit. Use standard Search performance and analytics data to evaluate traffic, while acknowledging attribution limitations. ### Mistake 2: Adding AI Impressions to Regular Search Impressions The dedicated AI reporting view represents exposure already incorporated into broader Search performance data. Adding the two totals produces an inflated number. ### Mistake 3: Assuming Every AI Impression Is an AI Overview The Search report covers both AI Overviews and AI Mode. The current dedicated report does not provide a separate feature-level breakdown. ### Mistake 4: Ignoring Canonicalization Unexpected URLs in the Pages report may result from canonical assignment or redirects. Review indexing and canonical signals before concluding that a particular URL received no exposure. ### Mistake 5: Interpreting Daily Fluctuations as Ranking Changes Google's reporting data can experience technical anomalies. For example, Google documented a Generative AI Search impression logging problem affecting August 13–17, 2026. The missing data was subsequently restored. The issue affected reporting rather than establishing an actual decline in search visibility. The [Search Console data anomalies page](https://support.google.com/webmasters/answer/6211453) is useful when investigating unexplained movements. ### Mistake 6: Confusing Search and Discover Metrics These reports cover different experiences and can use different aggregation methods. Their figures should not be treated as interchangeable measures of performance. ### Mistake 7: Claiming Proven AI Traffic Growth From Impression Growth Rising AI exposure is useful evidence of visibility, but it cannot independently establish additional clicks, sales, or subscriptions. A complete assessment requires broader traffic and business metrics. ## How Generative AI Reporting Changes SEO Strategy Traditional SEO reporting often starts with a familiar sequence: ```text Search query → Search impression → Website click → On-site engagement → Conversion ``` Generative AI introduces an additional discovery experience: ```text User question → AI-generated response → Supporting source link → Possible website visit → Engagement or conversion ``` The new report improves measurement around the supporting-source stage. It does not independently measure every subsequent outcome. This creates three distinct analytical priorities. **Visibility:** Determine whether relevant pages are appearing in supported AI experiences. **Engagement:** Evaluate whether broader organic traffic and on-site behavior improve alongside visibility. **Business value:** Assess whether visitors ultimately complete meaningful actions, such as using a tool, starting a trial, subscribing, or making a purchase. For publishers and directory websites, the most defensible approach is to track these stages separately instead of presenting AI impressions as equivalent to traffic or revenue. ## Frequently Asked Questions ### Is Generative AI Features (Beta) Available to Everyone? Google announced worldwide rollout completion on August 31, 2026. However, individual websites may not see a useful report if they lack sufficient qualifying AI impressions. ### Does the Report Include ChatGPT, Perplexity, or Other AI Search Engines? No. The Google Search Console report measures supported Google search experiences. It does not provide visibility data from independent AI products or competing search platforms. ### Can the Report Show Which Keywords Triggered AI Citations? No. The dedicated report does not currently provide a Queries dimension. The conventional Search results report contains query data, but it does not isolate the queries responsible for specific Generative AI appearances. ### Can Website Owners See AI Overview Clicks? Not separately in the dedicated Generative AI report. Click activity associated with supported Google Search experiences is included in broader Search performance reporting, without dedicated AI click totals in this view. ### Does a Higher AI Impression Count Mean Better Rankings? Not necessarily. The report does not provide a dedicated AI ranking-position metric. Higher impressions can reflect greater visibility, changes in search demand, or shifts in supported result presentation. ### Can Google AI Use a Website That Has Been Excluded From Generative AI Search Features? Google's exclusion control prevents eligible site content and links from appearing or grounding responses in the supported features covered by that setting. Google separately explains that the control does not govern all possible uses of content across Search or AI model training. ### Is Generative AI SEO Different From Traditional SEO? Google's official guidance presents AI search optimization as an extension of foundational SEO. Original content, crawlability, indexability, relevance, and good user experiences remain central. ### Should Websites Optimize Specifically for Google AI Mode? Websites should ensure their content answers real user needs comprehensively, but Google does not recommend creating large numbers of repetitive pages solely to target speculative AI-generated query variations. ### Can Generative AI Performance Data Predict Future Traffic? It can provide an early indication of discoverability, but it is not a reliable standalone forecasting model for clicks or conversions. Historical performance, audience behavior, search demand, and conversion metrics remain necessary. ## Conclusion **Google Search Console's Generative AI Features report marks an important improvement in AI search visibility measurement.** For the first time, website owners have a dedicated Google reporting view showing which pages appear in supported AI Overviews and AI Mode experiences, how often those appearances occur, and where their visibility originates. The 2026 rollout, multimodal reporting update, and improvements to source-link presentation reflect a broader transformation in how users discover information through Google Search. Yet the report remains intentionally limited. Without dedicated AI click data, query attribution, and conversion measurements, impressions should be interpreted as evidence of visibility rather than proof of commercial impact. The most effective strategy combines regular AI impression monitoring with established SEO practices, original content development, and independent measurement of meaningful website outcomes. **Open [Google Search Console](https://search.google.com/search-console), review the Generative AI Features report for your website, identify the pages gaining AI exposure, and use those findings to prioritize your next content and technical SEO improvements.**