Google Search Console 2026: How to Measure and Control Visibility in AI Overviews and AI Mode

For a long time, one question had no satisfactory answer: how can you tell whether your website is actually appearing in Google's AI-generated search experiences?

You could monitor referral traffic, run manual searches or rely on third-party tools, but there was no clean first-party view showing how often your pages surfaced in Google's generative search features.

That has now changed.

As of August 31, 2026, Google says it has completed the global rollout of new Search Console reporting for generative AI features in Search. These reports make it possible to track impressions generated through AI Overviews and AI Mode. Google has also expanded a control that lets site owners decide whether their content should remain eligible for these AI experiences.

For Swiss SMEs, this matters more than it may seem. It does not mean traditional SEO has suddenly become obsolete or that every company now needs a separate, mysterious "AI SEO" strategy.

In fact, Google's latest guidance points in the opposite direction: solid SEO fundamentals remain the foundation.

The real opportunity is more practical. Businesses can now create content that is useful enough to perform in both classic search and AI-powered experiences, then measure that visibility with first-party data instead of guesswork.

What Actually Changed in Google Search Console

Google announced dedicated generative AI performance reports for Search Console in June 2026, followed by a broader rollout that Google says became global on August 31, 2026.

The Search report can surface impressions from experiences including:

• AI Overviews, Google's AI-generated summaries displayed directly in search results;
• AI Mode, the more conversational search experience powered by generative AI.

Google also provides a separate report for generative features in Discover.

The important difference is attribution. Previously, appearances inside AI features were largely blended into general Search Console performance data. That made it difficult to understand whether a page was gaining visibility through standard search results or through Google's generative experiences.

The dedicated reporting creates a much clearer view.

According to Google, site owners can analyze metrics and dimensions such as:

• the number of impressions generated in AI experiences;
• the pages that appear most frequently;
• the countries where those impressions occur;
• the devices used for Search;
• performance over time across different date ranges.

For an SME, this is the first major step toward evaluating AI search visibility without relying purely on anecdotal testing.

Official source: Google Search Central – Generative AI performance reports

An AI Impression Is Not a Conversion

The new reports are useful, but there is an obvious trap: turning a new metric into a new vanity KPI.

An impression in AI Mode or an AI Overview means Google displayed a link to your site within a generative experience. It does not automatically mean someone clicked, requested a quote, booked a meeting or purchased anything.

At launch, the most actionable first-party signal is visibility, especially impressions.

To understand whether that visibility has commercial value, I would connect three layers of data:

1. AI visibility in Search Console: which pages are being surfaced?
2. Traffic and engagement: which visits actually reach the website?
3. Conversions: forms, calls, bookings, quote requests or sales.

This is the same principle that applies to traditional SEO. Visibility is not the end goal. The objective is to attract the right audience and give them a page that helps them move forward.

For a Geneva-based service company, for example, an article could accumulate AI impressions without generating any qualified business. Another page with fewer impressions might produce several high-value enquiries because its topic is closer to a buying decision.

That distinction matters when prioritising content.

Which Pages Should You Analyse First?

For most companies, the page-level report will probably be the most useful part of the new data.

Imagine a Swiss SME with twenty service pages and thirty blog articles. If only four pages repeatedly appear in Google's AI experiences, those pages deserve closer inspection.

They may share characteristics such as:

• answering a specific question clearly;
• providing expertise that is difficult to find elsewhere;
• using concrete examples rather than generic claims;
• presenting information in a well-structured format;
• being updated regularly;
• offering a complete answer without unnecessary filler.

The goal is not to copy their layout mechanically. It is to understand why these pages appear useful enough for Google to surface in a generative response.

The reverse analysis is equally valuable.

If an important commercial page has no measurable AI visibility, ask whether it is:

• too generic;
• too sales-oriented;
• almost identical to competitors' service pages;
• missing examples, evidence or local context;
• thin compared with the user's actual question;
• technically difficult to crawl or understand.

This turns "AI SEO" into a practical content improvement exercise rather than a speculative optimisation project.

Google Is Also Making One Point Clear: SEO Still Comes First

GEO and AEO discussions have sometimes created the impression that websites need to rebuild their entire search strategy to satisfy generative engines.

Google's official guidance is much more straightforward: the SEO best practices that already help Google understand and rank content remain relevant for AI features.

There is no separate magic markup required specifically for AI Mode.

The recommendations still look familiar:

• maintain a technically clear and crawlable website;
• publish original, genuinely useful content;
• provide a strong page experience;
• keep important pages indexable;
• use high-quality images and video when they add value;
• ensure structured data matches the visible content;
• demonstrate real experience and expertise rather than publishing generic content at scale.

This is important because the GEO market has attracted a lot of claims that are difficult to verify.

I previously covered the broader principles in our guide to GEO and AEO in 2026. Google's newer documentation now gives businesses a better basis for separating legitimate changes in search behaviour from tactics that are mainly repackaged marketing.

Google Explicitly Says Some Popular "GEO Hacks" Are Not Required

One of the most useful clarifications in Google's guidance concerns tactics that are often presented as essential for AI visibility.

Google says there is no need to rely on approaches such as:

• artificially breaking content into fragments only for AI systems;
• manufacturing mentions that are not authentic;
• creating extra text files such as llms.txt specifically to optimise for Google's AI search features.

That last point deserves nuance.

It does not mean llms.txt can never be useful in other contexts or for other systems. It means it should not be sold as a Google requirement, a ranking factor, or a shortcut to appearing in AI Mode.

For an SME, this is good news. Time and budget are usually better spent improving a genuinely valuable page than implementing a growing list of technical micro-tactics with no demonstrated impact.

Official source: Google guidance for AI features and your website

Search Console Now Also Gives Site Owners More Control Over AI Inclusion

Another important change is the addition of a specific control for generative AI experiences.

By default, eligible public website content may appear in features such as:

• AI Overviews;
• AI Mode;
• generative experiences in Google Discover.

A site owner can choose to opt out of these experiences.

Google explains that if content is excluded, the site will no longer be eligible to appear as a link in those generative features or contribute to grounding those answers. As a consequence, the site would also lose the related impressions and potential traffic.

For most SMEs, I would not disable this visibility by default.

If a public page can be surfaced as a source and potentially generate qualified visits, the opportunity is worth measuring before rejecting it.

Publishers, content platforms or organisations with specific distribution constraints may have different priorities and should evaluate the setting more carefully.

Official source: Search generative AI control – Search Console

What I Would Check Today on a Swiss SME Website

If you manage a business website, I would start with a simple five-step review.

1. Check Whether the Generative AI Report Is Available

Open Search Console and look for the performance reporting dedicated to generative AI.

Google says the rollout is global as of August 31, 2026. If you do not yet see meaningful data, that may simply mean the website has not generated enough impressions in these experiences.

Do not treat the absence of data as proof of a technical problem without further evidence.

2. Identify the Pages Already Appearing in AI Search

Sort pages by impressions and look for patterns.

Common candidates may include:

• in-depth guides;
• local landing pages;
• comparisons;
• specialist articles;
• product or service pages;
• case studies.

The objective is to identify editorial patterns, not just total impression counts.

3. Compare AI-Visible Pages With Your Commercial Priorities

Next, list the pages that matter most to your business and compare them with what is actually appearing.

If your strongest educational article is highly visible but your core service pages are absent, those service pages may need more depth, proof or differentiation.

For a local business, this can be particularly revealing. A generic "Our Services" page may have little informational value, while a detailed page addressing a specific Geneva or Swiss market problem could be far more useful to both users and search systems.

4. Fix the Technical Fundamentals Before Inventing an "AI SEO" Project

Before adding new AI-specific layers, make sure the website is already clean from a technical SEO perspective:

• indexing;
• canonical URLs;
• XML sitemap;
• site performance;
• structured data;
• internal linking;
• local business consistency;
• crawlability.

Our technical checklist for a high-performing Swiss website in 2026 covers these points in detail.

5. Measure Business Outcomes, Not Just AI Impressions

An increase in AI impressions can be encouraging, but it only becomes commercially meaningful when it contributes to real outcomes.

Track your important actions alongside Search Console data:

• contact form submissions;
• phone calls;
• consultation bookings;
• quote requests;
• purchases;
• other qualified conversions.

This prevents an AI visibility strategy from becoming disconnected from business performance.

Why This Matters for Companies in Geneva and French-Speaking Switzerland

This is not a US-only trend.

Google has already expanded AI search experiences into Switzerland, which means businesses in Geneva and across French-speaking Switzerland are competing in an environment where some searches are increasingly conversational and generative.

Users can ask more detailed questions, compare several criteria at once and expect a synthesised answer rather than ten blue links.

That can create opportunities for smaller companies.

A large brand does not automatically own every useful answer. A smaller business can still be surfaced if it publishes content that is more specific, more credible and more relevant to the exact question.

A generic article titled "Why Choose Our Company?" adds very little that Google cannot already infer from hundreds of similar pages.

A page explaining a real problem encountered by Geneva clients, with local constraints, realistic pricing, examples and first-hand experience, has much stronger informational value.

That is also how we approach content at Altio. The goal is not to publish articles simply to keep a blog active. The goal is to answer questions that potential clients genuinely ask before choosing a provider or investing in a website.

Should We Stop Using the Term GEO?

No, as long as we are clear about what it means.

GEO remains a useful label for the work involved in making brands and content more visible in generative search and AI assistants. The problem starts when it is treated as a completely separate discipline with its own set of unproven shortcuts.

Google's direction is increasingly clear:

strong SEO + genuinely differentiated content + machine-readable structure + specific measurement of AI visibility.

The major shift in 2026 is therefore not the arrival of a new "optimise for AI" button.

It is the ability to measure where content is being surfaced in Google's generative experiences and use that data to improve the broader search strategy.

Conclusion

Google Search Console's generative AI reporting is an important step because it moves part of the AI search discussion from speculation to measurement.

For a Swiss SME, the right response is not to rebuild the entire website for AI Mode.

I would start with three actions: open the new report, identify the pages already earning AI visibility, and compare their characteristics with the strategic pages that are not appearing yet.

Only then should you decide what needs to change.

SEO is not disappearing because search is becoming more generative. It is becoming more demanding about the things that have always mattered: technical accessibility, a clear match with search intent, credible expertise and content useful enough to deserve a citation.

Going Further

If you want to audit your own visibility before investing in new content, start with the fundamentals: our technical checklist for a high-performing Swiss website in 2026 lists what should already be in place before talking about AI.

• GEO, AEO and SEO in 2026 for a deeper look at the generative layer.
• Why your website is not generating clients if your site is visible but does not convert.
• How much does a website cost in Switzerland in 2026 to frame a realistic budget.
• Our website creation services and pricing if you would rather delegate the technical side.

Prefer concrete examples? Browse our portfolio, or start your brief: we will show you a prototype within 48h, already structured for both classic search and AI experiences.