Google added AI Search Visibility Reporting to Search Console, and where it chose to put it is the whole point. This is not a new product or a separate “generative console”. The same tool you already open to see your search ranking now also shows how often you show up in AI previews, AI Fashionand Discover’s AI capabilities. Where Google dropped the feature is Google settling a debate it’s been waging for a year: there isn’t a separate discipline called GEO, so there’s no separate place to measure it. AI visibility is search visibility, and it resides in the search tool.
If someone has spent the last year selling to you “generative engine optimization“As a new practice with its own playbook, its own budget line, and its own software subscription, Google has now expressed its disagreement in the most concrete way possible. It has told you what it thinks by pressing the button.
Reports show impressions, not clicks
Google’s announcement places new data in Search Console as generative reports on AI performance. They report impressions: how often your pages appeared in Google’s generative AI features in Search and Discover. You get the dimensions you already know from standard performance reporting, pages, countries, devices and dates, down to hourly granularity. It is being rolled out to a subset of UK websites first and then to a larger number.
Two absences matter more than everything in the report.
It doesn’t show clicks. Upon launch, you may see that you have appeared in an AI preview or AI mode response. You can’t see if anyone clicked, visited, or did anything. You get presence, not consequences. Google Search Console Help confirms impression reach only at launch.
It also arrives alongside a control that lets you opt out of your content from AI responses. One version gives you a counter for your presence in the AI responses and a switch to take you off. This association is indicative of Google’s position: it prefers to give you both the gauge and the output rather than continuing to respond to requests for one or the other.
Why location is the message
Search Console has been defining what counts as search performance for twenty years. What it reports is, by definition, what Google considers search performance. So when AI previews and AI mode impressions appear inside, next to your blue link impressions, that placement is an accounting decision: AI responses are search surfaces, and your visibility on them is search visibility.
Google has been making this argument for a while, that getting visible in AI search is the same job as being visible in search. I wrote about it when search and agents were consolidated into a single productand again when Google’s own guide to AI explained it. Integrating measurement into Search Console instead of creating a separate tool is this phrase compiled into software. Companies show you what they believe based on where they spend their engineering spend. Google spent it to rank the visibility of AI in search.
Free tool will bend where you look
As soon as Google’s AI visibility becomes free, native, and trackable in a tool that’s already open to every operator, it becomes the AI visibility that people are actually watching. Not because Google’s surface area counts more than that of ChatGPT, Claude or Perplexity. Because it’s the one with a free dashboard.
This is the street lighting effect applied to an entire channel. You look for your keys under the lamppost because that’s where the light is. Google is about to shine a bright, free light on its own surface, and the darker corners, the ones where standalone trackers charge you to look, attract less attention by default. The problem is that the visibility of AI is plural. Most pages cited by AI appear in a single enginetherefore a page cited constantly in one model may be absent in the next. A Google-only view is one engine of many, delivered to you with the authority of a number in a tool you already trust.
Multi-engine trackers are not the losers. They do the hardest thing Search Console will ever do, which is look into ChatGPT, Claude, Perplexity and the rest, where most of your AI visibility actually is. The risk is on the operator’s side, not his. It’s easy to put too much trust in a native, free, single-engine number, and “free and already in the tool I open every morning” can crowd out the number. transverse view that covers other surfaces. If anything, a free Google-only report makes multi-engine tools more necessary, because someone still needs to see the engines that Google will never report on.
And even within the Google surface, the number you are given is a leading indicator and not a result. Impressions indicate that you are present. They don’t say it mattered. This is exactly the trap of a recent episode on how AI visibility is confused with a business outcome: the metric is real, easy to track, easy to put into a deck, and disconnected from whether the buyer took action. A free impressions report is the most enticing version of this trap yetbecause the cost of removing it drops to zero.
Stop buying The GEO story is different
Takeaways are not a tool for learning. This is a story we need to stop believing.
Stop treating AI visibility as a separate workflow with its own budget and quarterly report. Google has now told you, by where it has placed the feature, that it is the same discipline measured in the same place. Integrate AI viewability tracking into SEO reporting pace you are already running. Pull the generative report for the Google slice when it hits your account. Keep a cross-engine check for surfaces that Google will never show you. Read each impression number the same way you read an impression in the standard report: a sign that you were eligible, not proof that you won. And make unsubscribing a deliberate decisionit’s not a flaw you’ll discover later.
The feature is useful. Placement is the point. Google ranked AI’s visibility in search because, for Google, that’s where it has always lived. Anyone who continues to pay for a separate GEO practice is optimizing a distinction that Google has stopped making.
More resources:
This article was originally published on No hacks.
Featured image: BestForBest/Shutterstock
