Your Google Analytics Is Hiding AI Traffic From You
Google AI Overviews and AI Mode send visitors to websites right now. GA4 records them as organic search. But those clicks never register in Google Search Console. Two tools, same user, contradictory data.
- 1When a user clicks a source link inside a Google AI Overview or AI Mode response, GA4 records the session as organic Google traffic. It is indistinguishable from a standard blue-link click.
- 2Those same AI Overview and AI Mode clicks do not register as clicks in Google Search Console at all. Your GSC click counts are understated, with no signal in the data pointing to the cause.
- 3Your organic traffic numbers in GA4 may already include a meaningful slice of AI-driven clicks you are attributing to rankings you did not earn through that channel.
- 4If your GSC impressions are holding steady while your CTR is falling on high-intent queries, an AI Overview absorbing clicks before they reach your result is the likely cause, but those clicks never appear in GSC to show you that.
- 5The gap between being ranked and being cited in AI is completely invisible in every standard analytics tool, including GA4, Google Search Console, Vercel Analytics, and Plausible.
Google AI Overview and AI Mode clicks register as organic Google traffic in GA4, but they do not register as clicks in Google Search Console at all. GA4 overstates what your rankings earned; GSC understates what your pages received. The two tools contradict each other, and neither will tell you why.
Open your GA4 account and look at your organic Google traffic for the last 90 days. Some of those sessions almost certainly came from a user who clicked a citation link inside a Google AI Overview, not from a traditional blue-link result. Now open Google Search Console and look at your click counts for the same period. Those AI Overview clicks are not in there: GA4 recorded the visit, but GSC never registered it.
This is not a theoretical future problem. Google AI Overviews now appear for approximately 16% of all tracked queries, reaching 2 billion monthly users globally. Each Overview includes roughly 13 source links. When a user clicks one, the session lands in GA4 as google / organic, identical to every other organic search visit. But GSC receives no click signal for that action. Your click counts in GSC are lower than your actual traffic, and the gap has no label.
GA4 counts the visit, but GSC never records the click. Same user, same action, two completely different numbers across your own tools.
What actually happens when Google AI sends you a click
Google AI Overviews generate citations in two ways: links shown alongside the summary text, and expandable source panels that list the pages Google pulled from to build the answer. A user can click either. When they do, the HTTP referrer is google.com, the medium is organic, and the session arrives in GA4 looking exactly like someone who found your result in the standard list and clicked through. GA4 records it. GSC does not register a click at all for that action.
Google AI Mode works the same way. AI Mode is a separate conversational tab where users ask complex, multi-step questions and receive long-form responses with citation links. Those citation clicks show up as organic Google sessions in GA4, but again produce no click event in GSC. There is no UTM parameter, no distinct referrer path, and no session property in GA4 that identifies the visit as AI-referred. And in GSC, the click simply does not exist.
The GA4 / GSC Split
Based on our testing: clicking a link inside Google AI Mode or an AI Overview registers a visitor in GA4 (as google / organic). It does not register a click in Google Search Console. This means your GSC click counts are systematically lower than your actual organic traffic, with no indicator explaining the gap.
The practical consequence is that any site appearing as a cited source in AI Overviews or AI Mode has an unknown but nonzero proportion of organic traffic that is actually AI-referred. For sites in categories where AI Overviews appear frequently (software comparisons, how-to content, service evaluations) that proportion could be substantial. There is currently no way to isolate it from the outside.
What Google Search Console does and does not show you
Google Search Console logs impressions from queries where an AI Overview appeared alongside the organic results. But clicks on AI Overview citation links do not register as clicks in GSC at all: the visit shows up in GA4 while Search Console receives nothing. This produces a concrete, testable discrepancy: your GSC click count for a page will be lower than the organic sessions GA4 records for the same page over the same period, with no explanation attached to the difference.
The CTR figure GSC calculates is therefore wrong in a specific direction. It divides clicks (which exclude AI Overview clicks) by impressions (which include the queries that triggered AI Overviews). The denominator is larger than the numerator implies. Your real click-through rate on those queries is higher than GSC shows, because some clicks are simply absent from the count.
58%
CTR drop for the #1 organic result when an AI Overview appears above it
Ahrefs, December 2025
16%
of all tracked queries now trigger a Google AI Overview
Serpstat, December 2025
0
filters in Google Search Console that isolate AI Overview impressions
Google, confirmed 2026
The most insidious version of this shows up when an AI Overview starts appearing for a query where you rank first. Two things happen simultaneously: users who click the Overview citation reach your site (GA4 counts them as organic), and users who read the AI summary and leave click nothing at all. GSC sees fewer clicks total and none of the AI-driven ones. Your position holds at one. Your CTR in GSC falls. The data looks like your content became less compelling, or the query intent shifted. The real cause, an AI Overview capturing engagement above your result, is not visible anywhere in the report.
GA4 is counting visits that GSC never recorded as clicks. The same user, the same action, two completely different numbers across your own tools.
The citation gap: ranked but not cited
There is a second problem layered on top of the attribution issue. Being cited inside a Google AI Overview is not the same as ranking for the query. Google selects which pages to pull from when building an Overview based on content quality, entity clarity, and how directly a page answers the specific question, not ranking position alone. A page at position five can be cited while the page at position one is skipped entirely.
The inverse is also true and more commercially damaging. A site that ranks well across its target queries but is never cited inside AI Overviews will experience steady CTR erosion as AI Overviews become more prevalent. Its rankings remain intact while its traffic quietly falls. GSC shows nothing obviously wrong. The cause (consistently absent from AI citations despite strong organic positions) does not appear anywhere in the standard measurement stack.
The Invisible Loss
If you rank well but are never cited in AI Overviews, you lose clicks to whatever is cited, but your rankings look fine and GSC shows no anomaly. The loss is unattributed. Standard analytics cannot surface it.
The sites that appear in AI Overview citations often see the opposite effect: CTR increases of 35% or more on cited queries, and research from ALM Corp and Semrush (2025) shows brands mentioned in AI responses experience 91% higher paid CTR on related terms as recognition compounds. The gap between cited and not cited widens over time, and the measurement tools most teams rely on will not show you which side you are on.
What you can actually detect, right now
Standard analytics cannot isolate AI traffic. There are still four signals worth monitoring that surface the problem indirectly, even without a dedicated AI tracking tool.
CTR decay on stable-impression queries
In GSC, filter for queries where impressions have held flat or grown over the past 90 days but CTR has declined. Sort by CTR change. These are your AI Overview exposure candidates. Go manually check each query in Google: if an AI Overview appears, that is the cause of the CTR drop.
LLM referral segments in GA4
Create a custom segment in GA4 for sessions where session source contains chatgpt.com, perplexity.ai, claude.ai, or gemini.google.com. This captures direct LLM referrals (clicks from standalone AI platforms outside of Google Search) which do carry a distinct referrer. Small volume, but higher conversion rates than organic average.
Manual AI Overview audit on target queries
Run your 20-30 highest-value queries in Google in a private window. Record which ones trigger an AI Overview, and whether your domain appears in the citations. This gives you a snapshot citation rate. It is not scalable, but it is the only direct measurement available without a third-party tool.
Position vs traffic correlation audit
Pull a list of queries where you rank in positions 1-5 in GSC. Cross-reference against actual organic landing page traffic for those pages. A consistent gap between expected traffic (based on position and impression volume) and actual traffic often indicates AI Overview presence absorbing clicks before they reach the organic list.
Google is one part of a larger measurement gap
The Google Analytics attribution problem is the most immediate version of this issue because Google is the largest search surface and the one most teams already instrument. But it sits inside a broader measurement gap that extends across every AI platform.
ChatGPT, Perplexity, Claude, and Gemini are all recommending products, services, and brands in response to buyer research queries. Some of that traffic reaches websites with a traceable referrer. Most of it does not: users copy a recommendation, type the brand name directly, or ask a follow-up question that eventually results in a direct or branded search session. The influence of AI on that session is invisible.
The platforms that generate AI recommendations draw from fundamentally different signals than Google does. Only 11-14% of domains cited by Google AI surfaces are also cited by ChatGPT or Perplexity (Averi, March 2026 analysis of 680 million citations). Ranking well in Google does not predict visibility in those platforms, and vice versa. Each channel requires monitoring on its own terms.
What GA4 / GSC can tell you
- ✓Organic click volume by query and page
- ✓Average ranking position in Google Search
- ✓Direct LLM referrals (chatgpt.com, perplexity.ai, etc.)
- ✓Impression volume by query
What GA4 / GSC cannot tell you
- ✗AI Overview or AI Mode clicks (GSC does not record them)
- ✗Whether an AI Overview appeared for a given query
- ✗Whether your brand was cited inside that Overview
- ✗Your citation rate in ChatGPT, Claude, or Perplexity
- ✗Which sources AI platforms are citing instead of you
What to do about it
The manual workarounds above give you directional signals. They do not give you systematic coverage. Manually checking 30 queries in Google every week does not scale to the hundreds of buying-intent queries that matter, and it tells you nothing about ChatGPT, Perplexity, or Claude.
BrandViz.AI tracks brand citation rates across ChatGPT, Claude, Gemini, and Perplexity, simulating real buyer queries at every stage of the purchase journey and recording which specific URLs each model cited in its response. Because Gemini shares the same underlying model family as Google AI Overviews and AI Mode, Gemini tracking provides a meaningful proxy for citation behavior on Google's AI surfaces.
The source-level attribution is the part that makes it actionable. Knowing your citation rate is one number. Knowing that a competitor is being cited because their G2 profile appears in 14 out of 25 vendor comparison queries, while your G2 page has 11 reviews and no structured comparison content, is a concrete fix list.
Start Without a Commitment
The free AI Visibility Snapshot from BrandViz.AI runs 25 real buyer queries through ChatGPT and returns results in roughly 10 minutes. It shows which queries your brand is absent from, which competitors appear instead, and which sources are driving those citations. No subscription required.
The GA4 / GSC discrepancy is not going to be fixed by Google in the near term. AI Overview and AI Mode clicks reach your site without sending a click event to Search Console, and arrive in GA4 with no signal distinguishing them from a standard organic visit. The more useful response is to measure AI citation rates directly, at the source, rather than waiting for downstream analytics to surface a problem the current data architecture cannot express.
Frequently asked questions
Can I create a GA4 segment to filter Google AI traffic?
Not for AI Overview or AI Mode traffic. Both surfaces send the same google / organic referrer as a standard blue-link click. GA4 records the session but has no property that identifies it as AI-referred. You can create segments for direct LLM referrals from standalone platforms like chatgpt.com or perplexity.ai, which do carry traceable referrer domains. For Google AI surfaces specifically, the session is indistinguishable in GA4, and the click does not appear in GSC at all.
Why do AI Overview clicks not show up in Google Search Console?
Based on our testing, clicking a link inside Google AI Mode or a Google AI Overview delivers the visitor to your site (GA4 records the session as google / organic) but does not send a click event to Google Search Console. GSC click counts reflect traditional organic result interactions only. Google has acknowledged the measurement gap and indicated that AI Overview reporting is on the roadmap, but as of May 2026 no GSC dimension or filter surfaces AI-driven clicks separately. Third-party tools that monitor AI citation rates directly are currently the only systematic alternative.
How much of my organic traffic might be from AI Overviews?
It depends heavily on the query categories your site targets. Google AI Overviews appear for approximately 16% of all tracked queries on average, but the rate is substantially higher for informational, comparison, and category-research queries, which is precisely where high-intent buyers are most active. Sites in software, professional services, and SaaS categories have reported AI Overview presence on 30-60% of their highest-volume informational queries. There is no universal figure, and your own numbers require a manual query audit to estimate.
Does good SEO mean I am already in AI Overviews?
Strong SEO is the foundation for AI Overview eligibility but does not guarantee citation. Google selects Overview sources based on how directly and completely a page answers the specific query being synthesised, along with entity clarity and structured data signals. A page ranked first can be skipped entirely while a page at position five is cited. The citation logic is distinct from the ranking logic, and the gap between them is what most teams currently have no way to measure.