For the past two years, AI search visibility has been the thing every marketer talks about and almost nobody can measure properly.
You know your content is being pulled into AI Overviews. You have probably seen your own brand quoted back at you in ChatGPT. But putting a number against any of it has been guesswork dressed up as reporting.
That changed on 3rd June 2026, when Google began rolling out dedicated generative AI performance reports in Search Console. UK site owners were in the first group to get access, which is unusual and genuinely useful if you run a UK business.
The reports are a real step forward. They also leave out most of what you would actually want to know. So here is what they show, what they hide, and how to build an AI search visibility setup that tells you something commercially useful.
TL;DR: Search Console now reports AI Overviews, AI Mode and generative AI features in Discover as separate reports. Impressions only. No clicks, no click-through rate, no average position, no query data, and no history before 18th May 2026. Treat them as an AI search visibility signal rather than a performance metric, then use GA4’s AI Assistant channel plus a custom channel group to work out what that visibility is actually worth.
What Google Has Actually Launched
Since mid-2025, appearances inside AI Mode and AI Overviews have counted towards your totals in the standard Search Console Performance report. The problem was that they were baked in. You could not separate them, so you could not tell whether a shift in impressions was traditional search moving or AI features moving.
The new reports split that out. There are three of them:
- AI Overviews in Search, covering the summary panels at the top of results
- AI Mode in Search, covering Google’s conversational search experience
- Generative AI features in Discover, reported separately from Search
Rollout started with a subset of UK-based sites, with wider availability planned but no firm date attached. If you have not got the reports yet, that is why.
What The New AI Search Visibility Reports Show You
One metric: impressions. How often your URLs appeared inside an AI feature.
You can break that down by:
- Pages, so you can see which URLs are being drawn on
- Countries
- Devices, for the Search reports
- Dates, with granularity from hourly through to monthly
Two things to know before you draw any conclusions. Data starts on 18th May 2026 and there is no historical backfill, so the day you first open the report is your baseline whether you like it or not. And Search Labs experiment data is excluded, so what you are seeing is live behaviour rather than tests.
What The Reports Don’t Show, And Why It Matters
This is the part most coverage glosses over. The gaps are significant:
- No clicks. You can see that you appeared. You cannot see whether anyone came through.
- No click-through rate. A direct consequence of the above.
- No average position. Every link inside an AI Overview shares a single position, so per-link optimisation has nothing to aim at.
- No query data. You cannot see what someone asked to trigger your appearance, which is the single most useful thing you could know.
Google has said metrics will be added over time. For now, the honest description is that these reports measure visibility, not value. That is still worth having. It is just not a substitute for knowing what AI search is doing to your pipeline.
How To Read Your AI Search Visibility Data Without Fooling Yourself
Because you have no year-on-year comparison and no click data, the temptation is to read patterns into noise. A few rules that keep this useful:
- Work at page level, not site level. The valuable insight is which pages AI features keep reaching for. That tells you what kind of content earns inclusion, and you can then write more of it.
- Read AI impressions alongside total impressions. If your overall impressions are flat but AI impressions are climbing, the mix of how you are being found is changing even though the headline number is not.
- Give it a quarter before you judge anything. With a May 2026 start date, short comparisons are close to meaningless.
- Don’t read a drop as a penalty. AI feature impressions move with how often Google chooses to show the feature at all, which is outside your control. We covered this pattern in more detail in why your impressions have dropped in Google Search Console.
Filling The Click Gap With GA4
Search Console tells you where you appeared. GA4 is where you find out whether that AI search visibility turned into anything worth having.
GA4 now has an AI Assistant default channel group. When Analytics recognises the referrer, it tags the session with a medium of ai-assistant and files it under that channel automatically, with no configuration needed from you. Google has named ChatGPT, Gemini and Claude as examples but has not published the full referrer list, which is precisely why you should not rely on it alone.
Two limitations you need to hold in your head when you look at those numbers:
- Referrer-less sessions still land in Direct. In-app browsers, mobile apps and copy-pasted links often strip the referrer, so a meaningful share of genuine AI traffic never gets attributed.
- AI Overviews and AI Mode clicks are not in this channel. They arrive as
google / organic, exactly like a normal blue link click. There is no native way to separate them in GA4, which means Search Console’s impressions and GA4’s clicks cannot be joined up for Google’s own AI surfaces.
Build A Custom AI Channel Group In GA4
This takes about ten minutes and gives you a consistent view that does not depend on Google’s undisclosed referrer list.
- Go to Admin > Data display > Channel groups.
- Create a new channel group and name it something obvious, such as AI Search.
- Add a new channel called AI Search, with the rule Source matches regex.
- Use a pattern covering the main assistants:
chatgpt\.com|chat\.openai\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|copilot\.microsoft\.com|deepseek\.com|grok\.com|meta\.ai|you\.com - Move the channel above Referral. GA4 evaluates rules top down, so if Referral comes first it will swallow the traffic.
- Save, then review it monthly and add new hostnames as assistants gain traction.
If you want to recover some of the Direct traffic, adding a UTM parameter to links you place yourself, and asking partners to do the same, closes part of the gap. It will never close all of it.
A Monthly AI Search Visibility Routine You Can Keep Up
Five things, once a month, and your AI search visibility stops being a mystery:
- Note total AI impressions from Search Console for AI Overviews and AI Mode.
- List your top ten pages by AI impressions and look for what they have in common.
- Pull AI Search sessions from your GA4 custom channel group, plus conversions from those sessions.
- Track Direct traffic as a control. A steady climb in Direct alongside rising AI impressions is a reasonable clue that unattributed AI traffic is growing.
- Record it in the same place every month. With no backfill available, your own record is the only long-run dataset you will ever have.
What This Should Change About Your Content
Measuring AI search visibility is not the goal, it is the feedback loop. Once you can see which pages AI features keep citing, the work becomes obvious: produce more of what gets picked up, and fix the pages that should be getting cited and are not.
In practice that usually means clear answers placed high on the page, specific claims rather than vague ones, and content structured so a machine can lift a single accurate paragraph out of it without mangling your meaning. If you want the fuller picture on that, our guides on how to rank in AI Overviews and traditional SEO versus AI SEO go further.
The businesses that get ahead here will not be the ones with the cleverest tactics. They will be the ones who started keeping a record in June 2026 while everyone else waited for better reporting.
Why can’t I see AI Overviews clicks in Search Console?
The generative AI performance reports currently include impressions only. Clicks, click-through rate, average position and query data are all absent. Clicks from AI Overviews do still count inside your overall Performance report totals, but there is no way to isolate them there, so you cannot calculate an AI-specific click-through rate. Google has said more metrics will follow.
Do AI Overviews clicks show as organic traffic in GA4?
Yes. Clicks from AI Overviews and AI Mode arrive as google / organic, identical to a standard organic click, and GA4’s AI Assistant channel does not capture them. That channel is for assistant referrals such as ChatGPT and Gemini. There is no native method for separating Google’s own AI surfaces from ordinary organic sessions.
Is a custom channel group still worth it now GA4 has an AI Assistant channel?
Yes, for two reasons. Google has not published the full list of referrers the native channel recognises, so you cannot tell what it is missing, and you control a custom group’s regex so you can add new assistants the moment they matter. Running both side by side also gives you a useful check: a large gap between them tells you the native list is behind.
How long before the new AI reports are actually useful?
Give it a full quarter of data before you make decisions about your AI search visibility on it. The reports start from 18th May 2026 with no historical backfill, which rules out year-on-year comparison for some time yet. In the meantime the page-level breakdown is immediately useful, because seeing which URLs AI features reach for does not depend on having a long trend line.
