Partly, and more than it could three months ago. Search Console now has a dedicated generative AI performance report, but it reports impressions only, with no queries, no clicks and no position, so it can tell you that something changed and never what was asked or what was said.
It is a genuine addition and it is not a visibility tool. Those are different statements and both are true.
Last updated: 30 August 2026. Search Console reporting for AI features is changing quickly, so this page is rechecked quarterly.
Most advice on this question is stale, including advice written earlier this year. The honest answer changed in June 2026 and will probably change again.
What actually changed?
Google added a generative AI performance report. Before that, appearances in AI features were folded into overall Search data with no way to isolate them, and the common advice, that Search Console could not tell you anything about AI surfaces, was correct.
It is now wrong. The report exists, it is separate, and per Google’s Search Console documentation it shows “how your site performs in generative AI features on Google Search,” including how “organic impressions from Search generative AI features changes over time” and “what pages are getting the highest (or lowest) impressions in generative AI features.”
What does it actually give you?
One metric and four dimensions. That is the whole surface, and knowing the shape of it is most of the answer to whether it is enough.
| Available | Not available | |
|---|---|---|
| Metrics | Impressions, defined as “how many times links to your site were shown to a user in a generative AI feature on Google Search” | Clicks, position |
| Dimensions | Pages, Countries, Dates, Devices | Queries |
| Granularity | Hourly, daily, weekly, monthly | Per surface, so AI Overviews and AI Mode are one bucket |
| Content | Nothing. It counts appearances | The answer text, and what was said about you |
The report draws on the same underlying data as the main report, since it “includes data from the Web search type in the Performance report (Search results).” So it is a filtered view rather than a new collection, which is why its dimensions are a subset rather than a superset.
What are the four gaps, and which one matters most?
- Consequence
- You can see a page gained impressions and never learn which question produced them. This is the big one.
- Consequence
- An impression counts whether you were the recommendation or the thing a competitor was described as better than.
- Consequence
- Nothing about ChatGPT, Claude, Perplexity, Grok, DeepSeek or Meta AI, which between them are most of the field.
- Consequence
- It counts links to your site. Being named in an answer with no link to you does not appear at all.
The fourth gap deserves particular attention, because it is easy to misread. This is a link-based measure, so it is closer to a citation count than to a mention count. A brand named in prose with no link is invisible to it, and that distinction is the one covered in mention rate vs citation rate.
How does it compare to what Microsoft gives you?
Unfavourably on the dimension that matters most, which is a genuinely surprising result.
Microsoft’s AI Performance report in Bing Webmaster Tools reports “the key phrases the AI used when retrieving content that was referenced in AI-generated answers,” alongside citation counts and page-level activity. Google’s report has no query dimension at all.
The smaller search engine tells you which questions retrieved you. The larger one tells you only that something did.
Both are free, both are first-party, and both are worth having configured. If you only have time to read one, read the one with the queries in it.
So what should you use it for?
Three jobs it does well, all of which are about direction rather than diagnosis.
Detecting that something moved. Impressions by date, at hourly granularity, is a good early-warning signal. A step change is real information even without the cause attached.
Finding which pages carry you. The Pages dimension tells you which URLs Google’s AI features actually surface, which is a usable input to where you spend content effort.
Geographic and device patterns. Countries and Devices are dimensions almost nothing else in this space gives you for free.
What do you still need something else for?
Anything involving a question or a sentence. Which prompts you appear for, what the answer said, whether you were recommended or merely listed, and how you compare to a named competitor on a specific question.
Also anything outside Google. Google’s own documentation on AI features notes that even its two surfaces differ from each other, since “AI Mode and AI Overviews may use different models and techniques, so the set of responses and links they show will vary.” Other vendors’ engines are separate systems again.
The practical arrangement is to use Search Console and Bing Webmaster Tools as your free first-party layer, then sample the engines directly for the things neither reports. That sampling is what a manual process or a tool both do, and the choice between them is priced out in how to track ChatGPT mentions manually and in build your own AI visibility tracker.
Recheck this answer in a quarter. Google shipped this report in June and it reports one metric today, which is exactly the shape of a feature that gets more dimensions later.