AI Search

How to Read Google's Generative AI Report in Search Console

Google now reports impressions from AI Overviews and AI Mode separately. Here's how to interpret the data, compare it with your usual search metrics, and avoid misleading conclusions.

Web Traffic Agents Editorial Team
Web Traffic Agents
·11 min read
Illustration showing search performance impressions connected to AI answer and traditional search result cards

Google Search Console can now show when pages from your site appear in Google's generative AI search experiences. That is valuable information, but it is easy to read too much into it. An impression is not a visit, and an increase in AI visibility is not proof that leads or revenue increased.

This guide explains how to find the new report, read it correctly, and turn its limited but useful data into a sensible editorial workflow.

What changed in Search Console?

On June 3, 2026, Google announced dedicated generative AI performance reports for Search and Discover. Google says the rollout reached websites worldwide by August 31, 2026. The Search report covers links appearing in AI Overviews and AI Mode. The separate Discover report concerns generative AI features in Google Discover.

This is a new view of visibility, not a replacement for ordinary Search performance reporting. Google's launch announcement says impressions from generative AI features remain part of its overall performance data.

That distinction matters because someone can see a link to your site in an AI-generated answer and never open the page. The new report helps you answer, "Which pages are being shown?" It does not answer, "How many customers did those appearances create?"

How to open the generative AI performance report

  1. Sign in to Google Search Console and choose the property you want to review.
  2. Look for the Generative AI report in the Performance area, or use Google's direct report link from the official help page.
  3. Set a date range. For a first review, choose a complete recent period, such as the previous 28 days.
  4. Start with the impressions trend, then open Pages to see the URLs receiving visibility.
  5. Use Countries, Devices, and Dates to look for patterns worth investigating. You can also review the supported Search type filter, which distinguishes text-based and multimodal searches.

If you cannot see the report, do not assume the feature is unavailable in your country. Google's documentation says that properties without enough generative AI impressions may not show it. Exclusion settings can also affect eligibility. A missing report does not establish a ranking problem.

The names and placement of interface controls can change, so rely on Google's help page if your dashboard looks slightly different.

What each metric or dimension actually tells you

ViewWhat it helps you understandWhat it does not establish
ImpressionsHow often links to your site were shown in supported AI search featuresVisits, clicks, or qualified leads
PagesWhich canonical URLs received reported AI-feature impressionsWhich exact prompt triggered a particular appearance
CountriesThe countries where the searches originatedCustomer location or buying intent
DevicesWhether impressions happened on desktop, mobile, or tabletConversion rate by device
DatesWhether visibility changed across days, weeks, or monthsWhy the change happened

An important technical nuance: Google says the main chart aggregates impressions by property, while a table grouped by Pages aggregates by URL. If your site appears with multiple URLs in one result, you may see different totals between those views. That is not necessarily missing data.

Google also notes that its most recent figures can be preliminary. Use complete date ranges for serious comparisons and avoid treating a one-day jump as an established trend.

A practical 15-minute review

You do not need an elaborate dashboard to learn from this report. Start with a small, repeatable routine.

Step 1: Record a baseline

Choose a complete recent period and write down the site's total reported generative AI impressions. Record the period and any filters you applied. An export of the chart and table is useful if you want a record before later updates.

Do not compare a partial week with a full month. For ongoing checks, compare similar date ranges and keep the same filters.

Step 2: Identify which pages appear

Open Pages and look at the highest-impression URLs. Make a short list of five to ten pages that matter to your business or publication. A tutorial with 2,000 AI impressions might deserve attention; so might a service page with only a few dozen if it represents a valuable query.

For each page, ask practical questions:

  • Does it answer a clear reader question early?
  • Is the information accurate and supported by primary sources?
  • Does it contain examples, original expertise, or useful detail beyond a generic summary?
  • Can a reader move naturally to a related guide, service, or next step?
  • Is the page indexable, canonicalized correctly, and discoverable through normal links?

An impression is evidence that a link appeared in an AI feature. It is not a score for authority or a guarantee of future exposure.

Step 3: Compare ordinary Search performance

Open Performance → Search results for the same property, pages, and dates. Review Web-search clicks and impressions there. Google includes AI Overviews and AI Mode in standard Web performance data, so do not add the two reports together as if they represented independent audiences.

The generative AI report gives you a specialized view of impressions. The standard Search performance report gives you broader metrics. Neither provides a reliable, stand-alone count of visits caused specifically by the AI-feature impressions in the new report.

Step 4: Look at on-site outcomes

Use your site's analytics to check engagement and conversions on the pages you identified. For a lead-generation website, that may mean calls, quote forms, or contact requests. For a publication, it could mean engaged sessions or newsletter signups.

Do not attribute those outcomes to AI Overviews simply because the same URL had AI impressions. Without more direct attribution, that relationship is only a hypothesis.

Our traffic intelligence guide explains why impressions, visits, and business outcomes should remain separate measurements.

Worked example: a guide gains AI impressions but not clicks

Imagine an educational guide with the following results. These figures are illustrative, not data from a real website.

MetricPrevious 28 daysRecent 28 days
Generative AI impressions for the guide1,2001,850
Total Web-search clicks for the guide310305
Contact or signup conversions from the guide, all channels1214

It would be fair to say that reported AI-feature visibility increased. It would be inaccurate to claim that AI traffic increased by 54% or that AI search produced two additional conversions.

A sensible next step is to check whether the guide still answers its subject well, whether its key facts are current, and whether visitors can find related resources. Then track the same metrics over another comparable period. The report gives you a clue about exposure, not a causal explanation.

What the new report cannot tell you

The initial generative AI performance report is intentionally narrower than many third-party AI visibility dashboards.

  • It is not a dedicated AI click or conversion report. The specialized report focuses on impressions.
  • It does not give you a dependable prompt-by-prompt ranking table. Don't infer precise positions or search prompts from the page list.
  • It does not cover ChatGPT, Claude, Perplexity, or other non-Google engines. Those are separate products with separate discovery and referral behavior.
  • It does not prove why Google selected a page. An impression is an observable appearance, not an explanation of a ranking or source-selection decision.

Be skeptical of any analysis that turns these impressions into a single "AI authority score" without showing how that score was calculated.

Search and Discover are separate reports

Google's generative AI Search report covers supported features in Google Search, currently including AI Overviews and AI Mode. Its separate Discover documentation describes how impressions work in generative AI features on Google Discover.

Do not blend the two into one metric without labeling them. Their surfaces and impression-counting rules differ. The Discover report, for example, explains that a link must come into view for a Discover impression and that scrolling past it twice in a session does not create two impressions for the same result.

If you publish evergreen how-to guides, the Search report may be the more natural place to begin. If Discover is important to your publication, analyze it separately.

What should you change on the website?

Usually, start with fundamentals instead of a new batch of "AI SEO" pages. Google's guidance for AI features says there are no additional technical requirements for appearing in AI Overviews and AI Mode beyond the existing Google Search requirements.

The most worthwhile checks are familiar:

  • Make sure important pages can be crawled and are eligible for indexing.
  • Put the primary answer in readable page text rather than hiding it in a client-only interface.
  • Keep titles, headings, internal links, and canonical URLs coherent.
  • Correct outdated claims and link to original documentation where relevant.
  • Use structured data when it accurately describes the visible page, not as a purported shortcut into AI results.
  • Avoid making a second article that repeats the same search intent just to target a different AI-related keyword.

The broader AI Search Visibility guide covers content structure and source clarity. For technical issues, the Technical SEO Audit is a better starting point than speculative AI-specific markup.

A reasonable reporting format for small teams

A one-page monthly summary is enough:

  1. Visibility: Total generative AI Search impressions, date range, and a comparison with the previous equivalent period.
  2. Pages: The five URLs with the strongest visibility and any meaningful changes.
  3. Search traffic: Overall Web-search clicks for those pages, clearly identified as broader Search data.
  4. Outcomes: Site conversions or engaged visits, without claiming they came from AI unless attribution supports that claim.
  5. Action: One or two substantive page improvements supported by the evidence.

Keep a note of changes you make. If an article receives more impressions next month, you will have context about what changed, though that alone still won't establish causation.

Frequently asked questions

Does the generative AI report show clicks from AI Overviews?

The dedicated report is an impression-focused view. Search Console's standard Web performance data includes activity from AI Overviews and AI Mode, but that does not make the dedicated report an AI-click attribution tool.

Why can't I see the Generative AI report?

Google says the report may be absent when a property does not have sufficient generative AI impressions. Check Google's current help documentation and your property's eligibility before treating absence as an error.

Do AI impressions mean Google cited my content as a source?

They indicate that links to URLs from your site were shown in supported generative AI features. They do not measure the quality of the citation, editorial trust, downstream clicks, or business value.

Should I write separate versions of every article for AI Overviews?

No. Google continues to recommend helpful, reliable content and ordinary SEO fundamentals. Improve the page that already serves the user's question before creating a duplicate.

Official sources and further reading

The practical takeaway: use the new report to learn where your content appears in Google's AI search experiences. Use ordinary Search Console metrics and website analytics to evaluate what people do next. Those are related questions, but they are not the same measurement.

Share

Written by

Web Traffic Agents Editorial Team

The Web Traffic Agents Editorial Team publishes practical guides on search visibility, AI discovery, analytics, content strategy, and conversion.

More from Web Traffic Agents Editorial Team →

Newsletter

Practical traffic intelligence in your inbox.

Occasional practical guides on search, AI discovery, analytics, or conversion. No tracking pixels in the body. Unsubscribe anytime.

Your address is never sold.

Related reading

Keep going