All of the Lab Journal
GEO / AI Visibility

How to See If AI Overviews Send You Any Traffic

Baptiste Lacroix
Founder of MentionLab
BlueWritten with Blue
July 7, 2026Updated July 15, 2026

Google does not label AI Overview clicks as their own channel. They land in your analytics as regular Google organic traffic, or as direct traffic with no referrer at all, so there is no single button that shows you "AI Overview visitors." As of June 3, 2026, Search Console has a new dedicated report that shows how often your pages appear inside AI Overviews and AI Mode, the same feature whose four-stage pipeline is broken down in how Google's AI Overviews actually work, though it only covers impressions, not clicks, and it is still rolling out to a limited set of sites. Until it reaches every account, four indirect signals in the data you already have can tell you almost as much.

Why Doesn't Google Analytics Just Show Me AI Overview Traffic?

Google Analytics and Search Console both fold AI Overview clicks into your normal organic search data, because Google does not attach a distinct referrer or channel tag to them. When someone clicks a link sitting inside an AI Overview panel, their browser typically still sends google.com as the referring domain, exactly like a click on a standard blue link, so GA4 records the session under Organic Search, source google, the same bucket as every other search visit. There is no separate row, no extra parameter, nothing that flags "this click came from the generative panel rather than the list below it."

The gap widens when someone does not click a link at all but copies a URL straight out of the summary text, which happens often on mobile where the AI Overview sits above the fold and the copy icon is one tap away. That visit arrives with no referrer whatsoever, and both GA4 and most other analytics platforms classify it as Direct traffic, indistinguishable from someone typing your domain from memory. Search Console has the same structural limit on its own Performance report: an AI Overview impression and any resulting click get counted in the exact same query and page rows as a regular organic listing, with no filter to split them out. That shared blind spot is the one part of this topic almost every write-up agrees on, and it is exactly what the new Search Console report was built to start closing.

What Changed on June 3, 2026: Search Console's New AI Performance Report

On June 3, 2026, Google added a dedicated Search Generative AI performance report inside Search Console, giving site owners their first native view of how often pages show up inside AI Overviews, AI Mode, and AI features in Discover (source: Google Search Central Blog, June 2026). The report breaks impressions down by page, by country, by device, and by date, using the same filtering interface as the standard Performance report, so anyone already comfortable in Search Console can pick it up without new training.

Two limits keep it from fully closing the measurement gap on its own. First, it currently reports impressions only, meaning it will tell you that a page appeared inside an AI Overview, but not whether that appearance produced a click, so it answers "am I visible" without answering "did that visibility pay off." Second, Google has confirmed the rollout is partial, reaching a limited set of sites rather than every Search Console property at once, so absence of the report in your account right now does not mean your pages are absent from AI Overviews, it may simply mean your property has not been included in the rollout yet. Both limits are exactly why the four signals below still matter even after this launch: they are the closest thing to a click-level view until Google extends the report further, and pairing them with the broader set of KPIs to track alongside AI Overview traffic gives a fuller picture than either data source alone.

What Are the 4 Signals That AI Overviews Are Touching Your Traffic?

Four patterns in data you likely already have point to AI Overview influence: impressions rising while clicks stay flat, a spike in unlabeled direct traffic, a stable ranking position paired with a falling click-through rate, and a query that visibly triggers an AI Overview when you search it yourself. None of these signals proves AI Overview involvement in isolation, but two or more showing up together on the same query is a strong enough pattern to act on.

The click-through-rate signal carries the most independent evidence behind it. Organic CTR on queries that display an AI Overview bottomed out at 1.3% in December 2025 before recovering to 2.4% by February 2026, an 85% rebound in two months (independent CTR study of 53 brands and 5.47 million tracked queries, April 2026). The same research found that being cited inside an AI Overview drives roughly 120% more organic clicks per impression than an uncited presence in that same AI Overview, yet a cited result still lands about 38% below the CTR a page would get with no AI Overview present at all. In practice, that means a falling CTR on a stable-position query is not automatically bad news, it may simply reflect where your page sits inside the AI Overview rather than a ranking problem to fix.

SignalWhere to check itWhat it likely means
Impressions climb, clicks stay flatSearch Console, Performance report, by queryYour page may be gaining AI Overview citations without gaining clicks
Unexplained rise in "Direct" trafficGA4, Traffic acquisition, by session source/mediumVisitors may be copy-pasting a link from ChatGPT, Perplexity, or Gemini (no referrer passed)
Average position holds steady, CTR dropsSearch Console, compare two date ranges on the same queryAn AI Overview is likely absorbing clicks without changing your rank
The query visibly shows an AI Overview when you search it yourselfManual search, logged out, matching your target countryConfirms the SERP feature exists for that exact query before you interpret your data

Once these four checks turn a vague suspicion into an actual pattern on a specific query, it becomes much easier to fold that signal into a running AI visibility score instead of treating each query as a one-off investigation.

This article, Blue could have written it for you: content optimized for Google + AI, without you writing a single word.

Try mentionLAB

How Do You Set Up a Simple Test to Catch AI Referral Clicks?

A basic UTM link test, and a custom channel group or regex filter in GA4, can catch some AI referral clicks without needing developer access to Google Tag Manager. Start with the test: build a link with a tag such as utm_source=chatgpt&utm_medium=referral, post it once wherever you would naturally share it, and confirm it shows up correctly in GA4's Traffic acquisition report under that exact source and medium. That single check tells you your UTM tagging works before you rely on it for anything ongoing.

For the continuous side, go to GA4's Traffic acquisition report, group by Session source/medium, and watch for chatgpt.com, perplexity.ai, claude.ai, and gemini.google.com as referring domains. Traffic from these domains has grown quickly enough to be worth watching closely: after ChatGPT began showing clickable brand links directly inside its answers on May 7, 2026, total referral traffic to tracked sites rose 157.7% in a single week, and traffic specifically to homepages rose 354.7%, pushing homepages from roughly 26 to 32% of ChatGPT referral traffic up to around 60% (independent clickstream research, updated May 2026). If your GA4 property shows any of those four domains as a source, that traffic is confirmed and countable, unlike the AI Overview clicks discussed above, which stay invisible by design.

Be honest about the ceiling here. No filter or UTM setup can guarantee full coverage of every AI-driven visit, since Google's own AI Overview clicks still carry no distinguishing tag no matter how you configure GA4. Every method available in 2026 narrows the blind spot, none removes it, and treating this as a shift from full certainty to partial visibility, rather than a fixable gap, keeps expectations realistic. Once you have earned visibility inside these systems, the next question is usually how to keep and grow it, which is where earning a citation once you know whether AI Overviews are costing you clicks and getting cited across ChatGPT, Perplexity, and Claude, not just Google become the natural next steps.

Should You Even Worry About Losing Traffic to AI Overviews?

Not every page is equally exposed: comparison-style and question-format queries trigger an AI Overview far more often than transactional ones, so the honest first question, covered in more depth in what AI visibility actually means, is whether your traffic even sits in the exposed zone. Comparison queries such as "X vs Y" trigger an AI Overview 95.4% of the time, question-phrased queries trigger one 85.9% of the time, while transactional queries trigger one only about 5% of the time (independent CTR study, April 2026).

That spread turns a vague worry into a workable checklist. If most of your traffic comes from bottom-funnel, transactional queries, comparison or CTR shifts elsewhere on the SERP matter far more to your business than an AI Overview does, since it rarely shows up on those searches in the first place. If a meaningful share of your traffic instead comes from comparison articles or FAQ-style pages, you sit squarely inside the exposed zone, and the four signals from the previous section deserve regular attention rather than a one-time check, the kind of recurring review covered in how to run an AI visibility audit. This kind of query-level thinking is part of the broader shift from optimizing purely for SEO to optimizing for how generative engines surface and cite content, where impressions and citations increasingly matter as much as raw rankings. Deciding which of those queries are worth continuing to chase and which are better left to AI answers is covered in more detail in which queries to concede to AI Overviews and which to defend.

Frequently Asked Questions

Does Google Search Console show clicks from AI Overviews? Not yet. The Search Generative AI performance report launched on June 3, 2026 reports impressions by page, country, device, and date, but Google has confirmed it does not include click data at this stage, and the rollout is still limited to a subset of sites (source: Google Search Central Blog, June 2026).

Why is my AI-referred traffic showing up as "Direct"? When a visitor copies a link straight out of an AI answer instead of clicking it, no referrer travels with that visit, so GA4 and most other analytics tools file it under Direct traffic by default. A rise in Direct sessions on pages you know appear in AI answers is one of the four signals worth checking, alongside a clean referring-domain check for chatgpt.com, perplexity.ai, claude.ai, and gemini.google.com in your traffic acquisition report.

Is a drop in click-through rate always bad news? No. Organic CTR on AI-Overview-triggering queries fell to a low of 1.3% in December 2025 before climbing back to 2.4% by February 2026 (independent CTR study, April 2026), and a page cited inside an AI Overview still earns roughly 120% more clicks than an uncited one on the same query. A falling CTR next to a stable ranking position often reflects your position inside the AI Overview rather than a ranking problem.

Do I need a paid tool to track AI Overview traffic? Not for the basics. Search Console, GA4, and a manual search of your target query cover the four signals in this article at no extra cost. The main limit is not budget, it is that Google still gives no reliable click-level tag to AI Overview visits, a gap even independently published research acknowledges, noting that only 2.8% of ChatGPT answers included a source citation as of August 2025, up from 0.6% in January 2025 (independent research tracking ChatGPT citation behavior, May 2026 update), a reminder that citation tracking across AI platforms is still an early and fast-moving discipline. Articles produced through MentionLab include built-in tracking of citations across ChatGPT, Perplexity, and Claude for exactly this reason, so this kind of check does not have to stay a manual habit.

Blue handles your SEO and your GEO. On autopilot.

You approve, she produces content optimized for Google + AI.

Join the Lab · 5-day trial

Read next

GEO / AI Visibility

Where ChatGPT Actually Gets Its Information

ChatGPT answers depend on two separate systems that work nothing alike: a training corpus frozen at a fixed cutoff, February 16, 2026 for the frontier models running today, and live web retrieval, which fetches current pages during the conversation itself, a capability every user has had since February 5, 2025. This article traces both paths using OpenAI's own 2026 documentation, including the specific crawlers behind each one, GPTBot for training and OAI-SearchBot for search, and the only training mix OpenAI has ever published in detail (GPT-3's, from 2020, not anything more recent). The second half turns practical: since a site cannot get into the training corpus on demand, it covers what actually makes a page retrievable and citable in ChatGPT's answers right now.

GEO / AI Visibility

What Zero-Click Search Means for Your Website Traffic

Zero-click searches end on the results page: the searcher gets an answer and never clicks through. Bain & Company (February 2025) puts the resulting drop in organic traffic at 15% to 25%, with around 60% of searches sending no visit to any third-party site. Pew Research Center (July 2025) and Google's own account of stable total click volume (August 2025) look like a contradiction; they are actually measuring different things, and both matter for reading your own Search Console data. The real work is triage: which queries to concede to the summary, which to defend with depth a summary cannot compress, and which to capture with both a citation and a reason to click.

GEO / AI Visibility

Where SEO Is Headed Over the Next Few Years

SEO is not disappearing, but the exchange behind it is being repriced by numbers that can actually be checked. Google users click a normal search result 8% of the time when an AI summary appears, versus 15% when it does not, according to Pew Research Center (July 2025). Meanwhile, Google's crawlers pull far fewer pages per referred visitor than AI answer engines do: about 5.4 pages per visitor for Google versus more than 38,000 for Anthropic, per Cloudflare data from July 2025. This article separates what has been measured from what is being projected, names the mechanism behind the shift, and reports the contrary evidence too, including that Google's own Search query volume just hit an all-time high.