How to Measure AI Overviews Traffic in Search Console and GA4

AI Overviews are quietly changing what your traffic looks like. A user asks a question, gets a summarized answer above the organic results, and often no longer clicks through. On your end, that looks like a page that suddenly underperforms, even though your position stayed exactly the same. The problem is that most companies only notice this once traffic has been leaking for months, because they do not know where to look.

In this article I show you how to concretely measure the impact of AI Overviews with Google Search Console and GA4, how to separate a real AI-driven click drop from noise, and which signals to track every month. This is a measurement guide, not a gut feeling. If you want to lay the foundation underneath all of this first, understand what SEO precisely covers and why organic positions remain the basis for both classic and AI visibility.

Why you do not simply see this in a dashboard

The first thing to understand: Google does not report AI Overviews as a separate channel. Impressions and clicks that come from an AI Overview are counted within your normal Search Console figures. There is no button that says “show me only AI Overview traffic”. That means the impact only becomes visible in the ratio between impressions, clicks and click-through rate, and not in a standalone report.

As a result, an AI-driven decline looks deceptively like other problems. A lower CTR can come from an AI Overview, but just as easily from a dropped position, a stronger competing result, a seasonal effect or a core update. Anyone who only looks at site-wide traffic sees a vaguely declining line and draws the wrong conclusion. The skill lies in isolating: on which queries and which pages does the CTR move while the position stays stable? That is the pattern that points to AI.

Step 1: lock in a baseline before you conclude anything

You cannot measure impact without a starting point. So begin by recording a reference period per query and per landing page. Open the Search results report in Search Console, switch on the three metrics that matter (impressions, clicks and average CTR) and export a period of preferably the past six months.

Split that export across two dimensions:

• By query: this shows you which specific searches are moving. Informational, question-driven queries (“what is”, “how”, “difference between”) trigger an AI Overview most often and are therefore your risk group.
• By page: this shows you which landing pages are losing their share of clicks, even when the underlying queries shift.

Keep this baseline outside Search Console, because the report by default only shows a limited history. A simple spreadsheet with the figures month by month is enough to demonstrate a trend later on. Without that starting point, every later drop remains a feeling rather than a measurement.

Step 2: read the deceptive pattern

Now for the core. The signal that points to AI Overviews is not “less traffic”, but a very specific combination:

• the impressions stay stable or rise (you still appear just as often, or more);
• the average position stays the same (you have not dropped);
• but the CTR falls and with it your clicks.

Those three together usually mean that something has come between your impression and the click, serving the user without them clicking through. An AI Overview is the most likely candidate for that, certainly on informational queries. If the drop instead appears together with a dropped position, you do not have an AI problem but a classic ranking problem, and the solution is a different one.

Carry out this analysis at the query and page level, never at the site level. The effect is heavily concentrated: part of your informational content can lose ground sharply while your transactional pages (“quote”, “price”, “[service name] agency”) are barely touched, because those queries produce an AI answer far less often. A site-wide average hides exactly the pages where it hurts.

Step 3: confirm in GA4 with landing page behavior

Search Console tells you what happens in the search results, GA4 tells you what happens next on your site. Use both together. Build a report in GA4 on the landing page dimension, filtered on the organic traffic channel, and put sessions and conversions next to it.

What you are looking for is confirmation of the pattern from step 2: the same pages that lose clicks in Search Console should show fewer organic sessions in GA4 over the same period. Watch for one important nuance effect here. If your traffic drops but your conversion rate rises, chances are that the disappearing visitors were mostly orientation-stage readers who were not going to buy anyway. AI Overviews often capture precisely that top, informational layer of the search query. That makes the drop less bad than it looks in absolute sessions, but you do have to measure it rather than hope.

If you want to dig into this more deeply, our guide on how to diagnose AI cannibalizing your organic traffic in GA4 and GSC helps, laying the same signals side by side step by step.

Step 4: rule out the alternative explanations

A measurement is only reliable once you have ruled out the other causes. So run through this checklist before you note down “AI Overviews” as your conclusion:

• Position: has your average position really stayed stable? If not, it is a ranking story.
• Season: compare with the same month last year, not just with the previous month. B2B traffic has predictable troughs around holiday periods.
• Core update: does the turnaround coincide with a known Google update? Then more than AI alone may be at play.
• Query type: is the drop on informational queries (AI-sensitive) or on transactional ones (less AI-sensitive)?

Only when the position is stable, the season does not explain it, and the drop is concentrated on informational queries do you have a substantiated suspicion of AI Overview impact. That discipline separates a real measurement from a panic reaction.

From measuring to acting

Measuring is not an end in itself. As soon as you know which pages are losing clicks to AI answers, you have two moves. First: protect the pages where the click still counts by making your content extractable, so that instead of being competed away you become the cited source. Second: shift your expectations on purely informational queries toward visibility and authority rather than pure clicks. How to trace where you have the best chance of appearing as a source in an AI Overview, you can read in our guide on finding the AI Overviews where you already rank.

This is exactly the work where an experienced SEO specialist makes the difference: not just reporting that traffic is dropping, but isolating the cause and reworking your content so that you remain the source in both the classic results and the AI answers. For us, SEO is not a standalone ranking exercise but the acquisition layer of one orchestrated growth engine, measured on pipeline and not on vanity positions.

Want to know how much traffic your informational pages are losing to AI Overviews and what you can do about it? Get in touch and we will read your Search Console and GA4 data together, with a concrete plan to turn your visibility into pipeline.