For a while, measuring AI search traffic felt like detective work without the interesting clues. People took screenshots of AI Overviews, checked referral reports, and tried to work out whether a mention had sent anyone to the site.
Google’s reporting is now more useful. In June 2026, Google launched a dedicated Generative AI performance report in Search Console for a subset of websites. It shows how often pages appear in generative AI features on Google Search, including AI Overviews and AI Mode. The rollout is gradual, so some properties still will not have it.
Here is the part I would put in bold for every marketing report: the dedicated report mainly measures visibility. It is not a complete traffic or revenue report. For that, you still need the ordinary Search performance report, Analytics, and your CRM.
Find The Report Before You Start Analysing
The first step is surprisingly easy to skip. Check whether the report is actually available in the Search Console property you are looking at.
A Missing Report Does Not Mean Zero Visibility
Google says the report may not appear when a site has not received enough impressions in generative AI features. It is also still being rolled out. So if you cannot see it, do not announce that Google never shows your pages in AI features. You simply may not have access to the report yet.
I would also check the property itself before blaming availability. A domain property, a URL-prefix property, a subdomain, and a country folder do not necessarily cover the same URLs. If your team has separate properties for www, non-www, a secure version, or a regional site, open the one that actually contains the pages you want to study.
Search Console also has a generative AI report for Discover. That is a separate product view. This article concerns Search, not Discover.
Understand What The Number Means
Google's report counts impressions from AI Overviews and AI Mode. It doesn't count website visits.
Impressions Are Not Traffic
Say your Search Console report shows 4,000 impressions. That number might look good in a client report, but I wouldn't make much of it on its own. How many people clicked through? Did anyone enquire about the service? That's what I'd check next.
A link can appear in an AI answer without bringing a single person to your site. People may get their answer there and leave. It happens, and the impression figure won't tell you otherwise.
The report can answer useful visibility questions:
Are our pages appearing in AI features?
Is exposure increasing or declining?
Which URLs are being shown?
Are impressions concentrated in a particular country or device?
It cannot, on its own, answer how many people visited, filled in a form, called the business, or bought something. Use the wider reporting stack for that.
Read The Report From The Top Down
My preference is to start with the chart and the date range, not the page table. Page-level numbers look exciting, but they can distract you from a seasonal spike or a very recent data wobble.
Build A Baseline First
Choose a period long enough to show a pattern. If the site has enough history, compare complete weeks or months. Looking at two unusual days and calling the difference a trend is one of the fastest ways to produce bad marketing commentary.
Google supports hourly, daily, weekly, and monthly granularity in the report. The dates use Pacific Time. That detail matters if your team works in the UAE, Asia, or Europe and compares Search Console with local analytics data. Midnight in Search Console may fall in the middle of the business day for your team.
The latest figures can also be preliminary. Google marks preliminary data with a dotted line and warns that it may change as more data is collected. I would leave the newest period out of a final monthly conclusion, or at least label it clearly.
A reasonable baseline includes the following:
The date the report first became available for the property
The last four complete weeks
A month-over-month comparison
A longer view if the site has enough history
If impressions rise but enquiries do not, that is not necessarily a tracking failure. The pages may be appearing for broad informational searches that do not carry buying intent. That is when I would move to the pages, queries, and conversion reports instead of celebrating the bigger number.
Use Pages To Find Where Visibility Lives
The Pages view shows which final URLs appeared in generative AI features.
Look For Concentration, Not Just The Leader
You might find that one blog post gets most of your AI impressions, while your service pages barely show up. I wouldn't worry about that straight away. Someone looking for advice is more likely to find a useful guide than a page asking them to buy something.
But I'd look at those pages more closely. Does the blog explain the topic better? Is it more up to date? Does it answer questions that your service pages leave out? Sometimes, a few gaps in the commercial pages explain the difference. Still, it tells you something about how Google sees the site. The article may contain clearer explanations, newer information, stronger examples, or more detail than the commercial pages.
Of course, it can work the other way around. Your product pages or local landing pages might get more impressions than your blog posts. Compare them and see what's different.
One small detail can also cause confusion when you're checking the numbers. Search Console groups page data by the final URL after redirects and generally assigns it to the canonical URL. If you've recently moved or redesigned your site, don't simply add the figures for old and new URLs together. You could end up counting data that Google has already grouped under one URL.
The Top Page Is Not Automatically The Most Valuable Page
The page with the most impressions may be a general explainer with almost no commercial intent. The page with fewer impressions may be a high-intent service page that produces actual leads.
I would export the page data and add a simple page-group label: service, product, editorial, location, support, or something similar. That small bit of housekeeping can show whether AI visibility is reaching the part of the site that matters to the business.
Compare Countries And Devices
The report can group impressions by country and device. This is where a global average starts becoming useful for real decisions.
A Country Filter Can Reveal The Local Story
Your pages might show up regularly in AI search results in the US but barely appear in the UAE. I've seen why looking at one overall number can be misleading. The difference might come down to local competition, language, how much people search for that topic, or whether your content answers what people in that market actually want to know.
If your business targets a specific country, check that market on its own. Look at AI impressions alongside regular Search impressions, clicks, and conversions. A big global number won't tell you much if the market you're trying to reach isn't responding.
There's another thing to watch if your site has both English and Arabic content. The combined figures can hide which language's pages are getting picked up. Check each group separately before deciding what's working.
Device Data Is A Clue, Not A Verdict
The Devices dimension covers desktop, tablet, and mobile. If impressions grow on mobile but engagement does not, inspect the mobile landing pages. Check the form, phone number, visible service details, and loading experience.
I would not conclude that mobile caused the problem from one device split. It tells you where to look, not what to fix.
Do Not Force The Chart And Table To Match
Search Console numbers do not always behave like a clean spreadsheet. That is normal here.
Why The Totals Can Disagree
You might notice that the chart and the table show different totals. That's not necessarily a mistake. Google counts impressions differently in each view. If two pages from your site appear in the same AI search feature, the chart may count that as one impression, while the page table lists the URLs separately.
So, adding up every row in the table won't always give you the same number as the chart. I've seen how easily this kind of mismatch can confuse a report, especially when someone expects every figure to line up. There's another catch. Search Console tables are limited to 1,000 rows. Exporting the data makes it easier to work with, but it won't bring back rows that aren't included.
My advice is simple: check the chart to see how performance changes, then use the table to find which pages, countries, or devices account for that activity. Don't expect both views to tell you exactly the same thing.
Pair AI Visibility With Business Data
The dedicated report tells you that links to your pages appeared. It does not tell you the commercial value of those appearances.
Keep The Normal Performance Report Open
Google says generative AI traffic is included in the overall Web search type in the ordinary Search performance report. That report remains the place to inspect clicks, impressions, CTR, average position, queries, and pages across Search.
The two reports will not create a perfect path from one AI impression to one sale. The Generative AI report is a visibility view. The standard Performance report is a broader Search view. Compare movement over time rather than trying to force the totals to match line by line.
For a monthly review, I would ask:
Did AI impressions rise for the page groups we care about?
Did total Web clicks rise, fall, or stay flat?
Did branded searches change?
Did organic leads or sales change in Analytics?
Did pages with AI visibility assist a conversion?
Analytics Tells You What Happened After The Click
In Google Analytics, look at landing pages, engaged sessions, form submissions, purchases, phone calls, bookings, and revenue. This assumes the events are configured properly, which is not a minor detail. Broken conversion tracking can make good SEO look useless.
Google has observed that clicks from Search results pages with AI Overviews can be higher quality on average, with users more likely to spend time on the site. That is an observation, not a guarantee for every business.
Your own conversion data should win the argument. I would take ten visits to a high-intent service page over a hundred visits to a general explainer if the ten visits produce qualified enquiries.
Build A Small Reporting Routine
You do not need a complicated data warehouse for the first few months. A spreadsheet is perfectly fine while the team learns what the report is actually telling you.
Keep The Fields Consistent
Record the reporting period, total AI impressions, top pages, leading countries, device split, total Web clicks, organic conversions, and revenue where available.
Add a notes column. Record a website migration, a new service, a product shortage, an algorithm update, a campaign, or a seasonal event. The AI report cannot explain those business changes for you.
Keep your page groups stable too. If “commercial pages” means service pages in January and product pages in April, the trend becomes fiction.
Add A Human Quality Check
Before celebrating a large increase, open a sample of the pages shown in the report. Confirm that each page is live, canonical, relevant, and says what you want associated with the business.
Visibility is not always flattering. A page may appear because it answers one narrow factual question while leaving out an important limitation. Check the whole landing experience, not just the number in the report.
Mistakes That Make The Report Look Better Than It Is
One mistake I see in SEO reporting is treating impressions as traffic. They aren't the same thing. Another is assuming the new report will show up in every Search Console account. Timing can trip you up, too. Comparing one day's unfinished data with a full month's results won't tell you much. I'd wait for the numbers to settle before drawing conclusions.
There's also no point adding the chart total to the page totals. Search Console counts them differently, so the figures won't necessarily match. Check the time zone settings as well. Search Console uses Pacific Time, while your Analytics reports may follow your local time. And if AI visibility goes up, don't rush to credit the whole marketing campaign. It's worth looking into, sure, but I'd want to see whether clicks, enquiries or sales improved before calling it a success. And do not treat an increase in AI visibility as proof that the marketing programme worked. It may be a promising sign. It still has to connect to useful visits, enquiries, transactions, or another outcome the business actually values.
Know What The Report Cannot Tell You
The report can show that URLs appeared in generative AI features. It cannot show every detail of how an answer was assembled, and it cannot guarantee that a customer saw exactly the same response you saw in a test.
Google’s AI features can use different models and techniques, and the links can vary. Search behaviour changes by country, device, context, and time.
I would treat the report as a strong new visibility signal, not a transcript of Google’s reasoning.
The practical setup is simple enough: use the Generative AI performance report for exposure, the ordinary Search Console Performance report for broader Search activity, and Analytics plus CRM data for business value.
That three-part view is much more useful than chasing one impressive AI visibility score.
FAQs
1. How Can I Track AI Overview Traffic in Google Search Console?
Open Google Search Console and look for the Generative AI performance report under Performance. It shows impressions from AI Overviews and AI Mode, broken down by pages, countries, devices and dates. For clicks and broader search traffic, use the standard Search performance report.
2. Can Google Search Console Track AI Mode Clicks?
The dedicated Generative AI performance report focuses on impressions, not direct click counts. However, clicks on links inside AI Mode are included in Google's standard Search performance reporting.
3. Why Are AI Overview Impressions Higher Than Website Visits?
An impression means your link appeared in an AI search result. A visit happens only when someone reaches your website. People may read Google's AI answer without clicking any source, so impressions can rise while website traffic stays flat.
4. Why Can't I See the Generative AI Performance Report in Search Console?
Your property may not have enough impressions from eligible AI features to qualify for the report. Check that you're viewing the correct Search Console property and that your site is allowed to appear in Google's generative AI search features.
5. How Do I Measure Conversions From AI Search Traffic?
Use Search Console to review AI impressions and overall search clicks. Then check Google Analytics for form submissions, purchases, bookings or other conversion events. If you need to measure qualified leads or revenue, connect those results with your CRM. Keep in mind that these reports don't automatically prove a conversion came from a specific AI appearance.
