Google Analytics 4 (GA4) does not have a ready-made report for AI traffic, but it records most of it. When someone clicks a link in an answer from ChatGPT, Perplexity, Gemini, Copilot or Claude, the visit usually arrives with the assistant's domain as its source, and GA4 files it under the Referral channel. You see AI traffic by filtering for those sources or by creating a custom channel group that collects them in one place.
How does GA4 know where a visit came from?
When a browser follows a link from one site to another, it usually sends a referrer, which is the address of the page the click came from. GA4 reads the referrer and stores the domain as the session source, for example chatgpt.com or perplexity.ai. The medium, a second label that describes the type of traffic, is set to referral.
A link can also carry UTM parameters, which are tags added to the end of a URL such as utm_source=chatgpt.com. GA4 gives these tags priority over the referrer. Some AI products add a utm_source tag to the links in their answers, so those visits are easier to identify even when the referrer is missing.
Visits from links inside AI answers are called AI referral traffic. In GA4 they look like any other referral until you group them.
How do you find AI sessions in the standard reports?
The quickest check uses the traffic acquisition report.
- Open Reports, then Acquisition, then Traffic acquisition.
- Change the primary dimension from Session default channel group to Session source / medium.
- Type a domain such as chatgpt or perplexity into the search box above the table, or add a filter on Session source.
The table then shows sessions, engaged sessions, conversions and revenue for that source. That is enough for a quick look, but you have to repeat the search for each assistant, and the results do not appear as a separate line in the channel overview.
For deeper analysis, use an Exploration (GA4's custom report builder). Put Session source in rows and add landing page as a second dimension to see which pages AI assistants send people to.
How do you build a custom channel group for AI traffic?
A channel group is the set of rules GA4 uses to sort sessions into buckets like Organic Search, Paid Search and Referral. You can copy the default group and add a channel for AI assistants.
Create the group
Go to Admin, then Data display, then Channel groups. Create a new channel group, which starts as a copy of the default rules.
Add the AI channel
Add a new channel, name it something like AI assistants, and set the condition to Source matches regex. A regular expression (regex) is a pattern that matches several text values at once. The pattern below covers the most common assistants:
chatgpt|openai|perplexity|claude\.ai|gemini\.google|copilot\.microsoft|deepseek|meta\.ai
Check the sources that actually appear in your own Session source report and add any AI domains the pattern misses.
Order the channels
GA4 assigns each session to the first channel whose rule matches. Move the AI channel above Referral; if you leave it below, Referral claims those sessions before your rule is checked. Save the group, then pick it from the dimension menu in the traffic acquisition report.
Custom channel groups apply to historical data as well, so the new channel shows past AI sessions too.
Which AI visits does GA4 miss?
Some AI traffic cannot be separated from other traffic in GA4, whatever filters you use.
| Situation | How it appears in GA4 |
|---|---|
| Clicks from a desktop or mobile app that sends no referrer | Direct |
| A user copies a URL from an answer and pastes it into the browser | Direct |
| Clicks on links in Google AI Overviews or AI Mode | Organic Search from google |
| An assistant reads your page through its own crawler | Not recorded at all |
| A user reads an answer that mentions your brand, then searches for it later | Organic Search or Direct |
Google's AI features sit on the same results page as the classic links, and the referrer is google.com in both cases. GA4 has no field that tells them apart, so look in Google Search Console for that traffic.
Crawler visits are missing for another reason. GA4 collects data through JavaScript that runs in the visitor's browser, and AI crawlers fetch the HTML without running that script. Server logs or a CDN dashboard show those requests, identified by user agent.
What does the GA4 number mean for a website?
The AI channel in GA4 counts clicks. It does not count how often an assistant read your page, quoted it or named your brand without linking. Because AI search writes the answer itself, many people get what they need without clicking, so the referral number is usually a small slice of the attention a site receives from AI answers.
Where GA4 helps is with behavior after the click. You can compare engagement rate and conversions for AI sessions against organic search and see which landing pages assistants send people to. Watching volume after you edit or publish content is also informative, and annotating those content changes makes the trends easier to read later.
To see the part that clicks do not show, sites track AI visibility: a fixed set of questions run across AI platforms on a schedule, recording when the brand is mentioned or cited. Searcherries does this with its AI visibility tracker. Read its results next to the GA4 AI channel to tell whether a drop in referrals comes from fewer citations or from fewer clicks on the same citations.
