Google’s New Generative AI Search Data: What SEO Teams Should Actually Do With It

For two years, SEO teams have argued about AI Overviews with almost no first-party evidence. A traffic dip appeared, one person blamed AI, another blamed a core update, and nobody could prove either theory.

That has partly changed. Google announced dedicated Search Generative AI performance reports in Search Console on June 3, 2026, and says they reached all websites worldwide on August 31, 2026. The reports show how often your pages appear in AI Overviews and AI Mode. They do not show how many people clicked.

This guide explains what Google’s new generative AI search data actually contains, where it can mislead you, and what an SEO team should do with it over the next 30 days, including whether the new opt-out control deserves a place in your plan.

Quick Answer: What Should SEO Teams Do With Google’s Generative AI Search Data?

Use the new Search Console report as a visibility baseline, not a traffic report. It counts impressions inside AI Overviews and AI Mode, grouped by page, country, date and device. That is enough to show which of your pages Google’s AI features draw from, how large that exposure is, and whether it is growing or shrinking.

It is not enough to prove revenue impact, because it contains no click data. The sensible playbook is short:

Measure: track AI impressions monthly and compare them with your regular Web search impressions to get a rough AI-exposure share.
Diagnose: find the pages Google’s AI features already retrieve, and ask whether they deserve to be there.
Improve: strengthen unique, first-hand content and fix any technical eligibility problems.
Decide carefully: treat the new opt-out control as a business decision that needs a baseline, not a reflex.
Ignore the noise: skip llms.txt files, content chunking and other “AI hacks” that Google says are unnecessary for its own AI features.

What Google Actually Released

Two things arrived together on June 3, 2026, and it helps to keep them separate in your head.

The first is a pair of performance reports, one for Google Search and one for Discover. They give site owners dedicated views of impressions inside generative AI features: AI Overviews and AI Mode in Search, plus generative AI features in Discover. The second is a Search Console control that lets a site choose whether its links and content appear in, and help ground, those features.

Google also published a guide to optimizing for generative AI features. It matters as much as the data does, because it tells you what Google itself says works.

The scale explains why Google built all this. By Google’s own figures, AI Overviews has more than 2.5 billion monthly active users and AI Mode has passed one billion monthly users. If a share of your audience meets your brand inside those answers first, you need a way to see it.

Rollout timeline

Date	What happened
May 2026	Google publishes a dedicated guide to optimizing for generative AI features (the page now shows a July 10, 2026 update).
June 3, 2026	Google announces the Search and Discover generative AI performance reports plus the Search generative AI control, testing with a subset of UK site owners.
June 3, 2026	The UK Competition and Markets Authority issues a conduct requirement on AI search opt-outs, according to Search Engine Journal.
July 2026	The control starts appearing on accounts outside the UK, per Search Engine Journal.
August 31, 2026	Google says the reports and control have rolled out to all websites worldwide.
March 2027	CMA deadline for Google to introduce page-level controls, according to Search Engine Journal.

What each report covers

The Search report covers AI Overviews and AI Mode, and Google says it expects to update that list over time. A separate report for Discover covers generative AI features there. In the Search version you can group data by page, country, date and device. Impressions mean the number of times links to your site were shown to a user inside a generative AI feature; Google explains the counting rules in its article on how impressions are counted.

Two details are easy to miss. Data from Search Labs experiments is excluded, and sites without enough AI impressions may not see the report at all, as the Search Console help page explains.

What the Data Includes and What It Leaves Out

Included	Not included (as of September 2026)
Impressions in AI Overviews and AI Mode	Clicks from those features
Grouping by page, country, date and device	Click-through rate and average position
Daily, weekly and monthly date views	A documented split between AI Overviews and AI Mode
Export of chart and table data	Search Labs experiments
A separate report for Discover	Full page lists on large sites (1,000-row limit applies)

Google’s documentation describes impressions only, and Search Engine Journal’s coverage of the worldwide rollout confirms there is no click data yet. Third-party analysis of the report makes the same point about click-through rate and position.

The AI numbers are a slice of your existing impressions

This is the most common reporting mistake. Google says the generative AI report includes data from the Web search type in the standard Performance report, and its announcement notes that the same data stays in the overall performance report. In other words, the new report separates out impressions you were already receiving. It does not create new ones.

A simple illustration: if your Web search impressions were 250,000 last month and the AI report shows 40,000, your total is still 250,000. Roughly 16% of that exposure happened inside AI features. Adding the two figures would invent 40,000 impressions that never existed, and the error grows as AI features take a larger share.

Why totals may not match

Aggregation differs by view. In the chart, two links from your site inside one AI response count as a single property-level impression, while the page table counts each page on its own. Google says chart and table totals can differ for that reason. The usual Performance report limits apply too, including the 1,000-row cap and Pacific Time dates, and the newest data can be preliminary.

Why Impressions Without Clicks Still Change Your Strategy

A fair objection: impressions alone do not pay invoices. Search Engine Journal is right that the reports don’t show how many clicks an appearance generates. Even so, impressions answer questions you could not answer before.

Which pages does Google retrieve? Google’s guide explains that its AI features use retrieval-augmented generation, pulling relevant pages from the Search index before composing a response. Page-level data shows what got pulled.
Which sub-questions do you match? Google also uses query fan-out, running several related searches behind a single prompt. Imagine someone asking for the best CRM for a small law firm. Hidden searches might cover client intake forms, per-user pricing and data security. If your pricing page surfaces, the report shows it, even though you never optimised for that prompt.
How big is the exposure? Without a baseline, an opt-out decision or a forecast is guesswork.
Where should you watch for click changes? Pages with rising AI impressions become a watch list inside the regular Performance report.

Think of the report as a map, not a scoreboard. It tells you where to look. It does not tell you what you earned.

How to Read the Report: A Six-Step Workflow

Step 1: Confirm access and eligibility

Open the Generative AI performance report in Search Console. If it is missing, Google’s help page lists three likely causes: your property has not received the rollout yet, your site has not had enough AI impressions, or you have excluded the site through the new control.

Eligibility is a separate question. According to Google’s guide, a page must be indexed and eligible to appear in Search with a snippet, meeting the Search technical requirements, and the site must be included in the generative AI features setting. Meeting the requirements still does not guarantee crawling, indexing or serving.

Step 2: Set a baseline and export it

Choose a full-month window, note the totals, and download the chart and table with the export button. Because standard Performance report limits apply, treat monthly exports as your own archive. If Google adds clicks later, you will want clean historical impressions to compare against.

Step 3: Rank pages and classify them

Sort by pages and copy the top 50 to 100 URLs into a sheet. Add the columns below, then tag each URL by content type: guide, comparison, service page, product page or local page.

Column	Where it comes from	Why it matters
URL	Generative AI report, Pages tab	Shows which pages Google’s AI features link to
AI impressions	Generative AI report	Measures the size of AI exposure
Web search impressions	Standard Performance report, same page filter	Gives you the denominator for AI-exposure share
Clicks and CTR	Standard Performance report	Lets you watch for divergence over time
Content type, last update	Your CMS	Shows which formats surface most often on your site
Owner, next action	Your team	Turns data into a backlog

Step 4: Split by country and device

Group by countries, then by devices. Businesses selling in several markets should look for uneven exposure. A page might surface often for searchers in one country and rarely in another, which can point to a localisation or intent gap. Device splits show where your audience meets AI answers, which should inform page design and mobile experience. Look for gaps first and avoid assuming a cause.

Step 5: Compare against classic performance data

Filter the standard Performance report to the same pages. If a page’s AI impressions climb while its click-through rate slips, you have a hypothesis: some searchers may be getting what they need from the answer. It is a hypothesis, not proof, because seasonality, ranking changes and layout shifts can produce the same pattern. Use custom annotations to log content changes and Google updates so you do not credit or blame the wrong event.

Step 6: Turn findings into a prioritized backlog

Sort your list into three buckets: pages that surface and deserve more investment, pages that surface but are thin or outdated, and important pages that do not surface at all. Each bucket needs a different fix, which the next section covers.

What SEO Teams Should Actually Do: Five Moves That Hold Up

Google’s own guide makes a point that should calm the room: its generative AI features are rooted in core Search ranking and quality systems, so established SEO practice remains relevant. That is why these recommendations lean on fundamentals rather than novelty. For wider context, we have covered how AI search is reshaping traditional SEO separately.

Move 1: Publish one AI-visibility metric every month

Define a single, directional metric: AI impressions divided by Web search impressions for the same period and property. Track it beside your usual traffic and conversion numbers, label it clearly, and never present it as AI traffic. It measures exposure, and exposure is all this data can prove. If you are rebuilding your reporting stack, our guide to SEO KPIs that actually prove results shows how to keep visibility metrics and business outcomes in the same conversation.

Move 2: Strengthen the pages that already surface

Google’s guidance says unique, non-commodity content will probably influence AI visibility more than any other tactic. Commodity content restates what anyone could write. Non-commodity content adds first-hand experience, original data or expert judgment. Compare a generic list of meta description tips with a write-up of what actually happened to click-through rates when one specific site rewrote its meta descriptions. The second gives a reader, and a retrieval system, something the first cannot.

For every page in your top bucket, ask five questions. Who wrote this, and why are they qualified? What first-hand evidence is on the page? Are claims sourced to primary references? When was it last reviewed? What would a reader learn here that they could not learn from ten other pages? These are the practical expression of the experience, expertise, authoritativeness and trust signals behind Google’s helpful, reliable, people-first content guidance. Our SEO and content strategy work starts from the same audit.

Move 3: Protect technical eligibility and business data

Nothing in the new data changes the basics: pages must be crawlable, indexable and snippet-eligible. Audit templates for robots directives, noindex tags and restrictive snippet settings that could quietly keep pages out; Google’s snippet controls documentation explains how those directives work. If you rely on JavaScript, confirm that key content renders properly by following Google’s JavaScript SEO basics, and check that pages meet page experience expectations on every device.

If you run a local or ecommerce business, Google’s guide also points to Merchant Center feeds and Google Business Profile as ways for products and services to appear in AI responses. Keeping those accurate is cheap insurance. When template or rendering fixes need engineering time, a development team that understands SEO saves several rounds of back-and-forth.

Move 4: Answer the whole question instead of chasing variations

Fan-out makes it tempting to publish a separate page for every imaginable sub-question. Google warns against that: creating pages mainly to manipulate rankings or AI responses can violate its scaled content abuse policy, and a large page count does not make a site better. The stronger approach is one comprehensive page that answers the primary question and its obvious follow-ups in clearly headed sections. That structure helps readers first and retrieval second, which is the right order of priorities. Our explainer on what GEO means in practice for Indian businesses adds context on the wider AI-assistant landscape.

Move 5: Balance organic exposure with channels you control

If AI impressions rise on informational pages while clicks stay flat, do not assume the sky is falling, and do not assume it is not. Reduce dependence on any single surface. Send high-intent visitors to conversion-ready pages, build email capture into your best content, and use performance marketing to protect lead flow in segments where organic click-through is under pressure. Where search intent is commercial and immediate, well-structured Google Ads campaigns can capture demand that a generated answer might otherwise satisfy.

Should You Use the New Opt-Out Control?

The Search generative AI control lives in Search Console settings and works at property level. Sites that opt out receive no traffic or impressions from AI Overviews, AI Mode or AI Overviews in Discover. Google says the setting is not used as a ranking signal for results outside those features, and Search Engine Journal notes it has no effect on AI training, which the separate Google-Extended control governs. Google’s help page describes how to exclude your site and how to include it again.

Regulation is part of the backdrop. According to Search Engine Journal’s reporting, the UK’s Competition and Markets Authority required Google to let sites opt out of AI search features without being penalized in standard results. The regulator has also set a March 2027 timeline for page-level controls.

There is no universal answer. These questions help frame it:

Question	What the answer suggests
Does most of your revenue come from page views or ad impressions?	Model the opt-out seriously. AI answers may substitute for visits.
Do you sell services or products and rely on brand discovery?	Staying in often makes sense, because appearing in answers can work as top-of-funnel exposure.
Is your content proprietary, licensed or paywalled?	Review your licensing position. Opting out may protect its value.
Do you lack baseline data?	Wait. Collect at least two months of impressions before deciding.
Might page-level controls arrive soon?	Consider waiting. A finer control may make a site-wide switch unnecessary.

This is a business decision, and the answer can differ by property. Model the downside before flipping a site-wide switch, and do not assume any change takes effect instantly.

What to Ignore: AEO and GEO Myths Google Has Addressed

The internet is full of AEO and GEO advice. Google’s guide is unusually direct about which popular tactics its own Search features do not need. From Google’s perspective, optimizing for generative AI search is optimizing for the search experience, which is still SEO.

llms.txt files and special markup. Google Search does not use them. Creating one will neither help nor harm your Google visibility, though it is fine if another service you care about reads it.
Chunking content into tiny pieces. Google says its systems can understand multiple topics on one page, and there is no ideal page length.
Rewriting copy just for AI. The systems understand synonyms and meaning, so you do not need every long-tail variant.
Chasing inauthentic mentions. Manufactured mentions across the web are not as helpful as they look.
Over-focusing on structured data. It is not required for generative AI search, though it remains useful for rich results.

Google also cautions that no third-party tool has access to its internal ranking or AI systems, so claims of secret Google metrics deserve scepticism. Read its guidance on evaluating third-party SEO advice before you buy a tool.

One nuance: this guidance covers Google Search. Other assistants such as ChatGPT and Perplexity have their own retrieval behaviour, so broader generative engine optimization work can still be worthwhile. Just do not confuse it with what Google requires.

How to Report AI Visibility to Clients and Leadership

Executives hear “AI Overviews” and assume disaster. Clear reporting language prevents overreaction.

Say “impressions in AI features,” not “AI traffic.” The data contains no clicks.
Show a trend and a share. Report monthly AI impressions and the AI-exposure share of Web search impressions, with a note that the share is directional.
List the top ten AI-exposed pages and the action planned for each.
Pair visibility with outcomes. Put leads, revenue or enquiries next to the impressions so nobody mistakes exposure for results.
Log changes. Annotate content updates and platform changes so movements have context.

What Comes Next, and What Is Not Confirmed

Three developments are worth watching, and they carry different levels of certainty.

More metrics. Google says it will introduce additional metrics over time and is working with site owners on which insights help most. No date has been confirmed for click data.
Page-level controls. According to Search Engine Journal, the CMA has given Google until March 2027 to introduce them, and it also expects click-through data.
Compliance reporting. Per Search Engine Journal, Google must report on its compliance every six months during the first year.

These are regulatory and product timelines, so they can move. Check the Google Search Central blog before making decisions that depend on them, and keep exporting your impression data so you can compare it with clicks when they arrive.

A 30-Day Action Plan for SEO Teams

Week	Focus	Deliverable
Week 1	Access, eligibility and baseline	Confirmed report access, first monthly export, snapshot of AI impressions by page, country and device
Week 2	Classification and comparison	Sheet joining AI impressions with Web impressions, clicks and CTR for the top URLs
Week 3	Fixes on priority pages	Content upgrades on top pages, technical eligibility audit, business data checks
Week 4	Reporting and decisions	Monthly reporting template, annotated timeline, written opt-out recommendation

Frequently Asked Questions

Does Search Console show clicks from AI Overviews and AI Mode?

Not in the generative AI performance report. Google’s documentation describes impressions grouped by page, country, date and device, and Search Engine Journal confirms click data is not included yet. Google says more metrics will arrive over time.

Are AI Overviews and AI Mode reported separately?

Google’s help page lists both as capabilities included in the Search report, but it documents no dimension that separates them. Assume a combined view unless Google updates the report.

Why can’t I see the report in my property?

Google lists three causes: the rollout has not reached your property, your site has not received enough impressions in generative AI features, or you have excluded your site through the new control. Check the setting first, then wait and look again.

Should I add AI impressions to my normal impressions?

No. AI impressions are already counted in your Web search totals, so adding them double counts. Use them as a slice of existing exposure instead.

Will opting out hurt my rankings in regular Google Search?

Google says the control is not used as a ranking signal outside generative AI features. The trade-off is direct, though: you receive no traffic or impressions from AI Overviews, AI Mode or AI Overviews in Discover.

Do I need llms.txt, special schema or chunked content to appear in AI Overviews?

No. Google’s guide says Search does not use llms.txt, does not require chunking, and needs no special markup for its generative AI features. Focus on unique content, technical clarity and page experience.

Is traditional SEO still relevant?

Yes. Google says its generative AI features are rooted in core Search ranking and quality systems, so foundational SEO practices still apply. What changes is measurement, plus a heavier emphasis on non-commodity content.

How often should we check the report?

Monthly is enough for most teams, with weekly checks around major content launches or site changes. Export regularly, because the standard Performance report limits apply.

Key Takeaways

Google’s generative AI performance reports show impressions in AI Overviews and AI Mode, not clicks.
The AI impressions are a subset of your Web search impressions, so never add the two together.
Use the data to build a baseline, find the pages Google retrieves, and prioritize content and technical fixes.
Unique, non-commodity content and clean technical eligibility matter more than llms.txt, chunking or other AI hacks.
Decide on the opt-out control with baseline data and a clear view of your revenue model, and revisit it as Google adds metrics and page-level controls.

If you would like help turning Search Console’s AI data into a practical roadmap, talk to an SEO strategist at Pixels Corp, or look at the work we have done for clients first.