Traffic from ChatGPT and Perplexity hides in your referrals. Build a custom GA4 channel group to isolate AI-referred sessions in four steps and track the trend.
Traffic from AI engines like ChatGPT and Perplexity arrives as referrals, but GA4 lumps it into generic buckets unless you tell it otherwise. By building a custom channel group that isolates AI-referred sessions, you can see how much traffic AI sends you and watch the trend over time. This guide walks through the setup in four steps.
When someone clicks a link inside an AI answer, they land on your site with a referrer like chatgpt.com or perplexity.ai. GA4 does not have a built-in "AI" channel, so those sessions usually fall into Referral or, if the link lacks a clear referrer, Direct.
The result is that AI-driven visits blend in with everything else, and you cannot tell how much of your traffic AI is actually sending. A custom channel group fixes that by naming the AI sources explicitly.
Keep in mind that much AI influence never produces a click at all — this method captures the visible portion, which is still worth measuring.
Start by collecting the referral domains AI engines use when they link out. Common ones include chatgpt.com, perplexity.ai, and the domains of other assistants your audience uses.
Check your existing referral report for these hostnames to confirm which are already sending you traffic. That gives you a real starting list rather than a guess.
Keep the list somewhere you can update it, since new engines and domains appear over time.
In GA4's admin, create a custom channel group so you do not disturb the default one. This is where you will define an AI channel.
Custom channel groups let you re-classify sessions by your own rules while leaving GA4's standard reporting intact. Name the group clearly so anyone reading your reports understands it.
Inside the custom channel group, add a channel — call it something like "AI Assistants" — and define a rule that matches your AI referrer domains. Use a condition on the source that matches chatgpt.com, perplexity.ai, and the others on your list.
Placing the AI channel before the catch-all Referral rule ensures AI sessions get labeled as AI rather than falling into generic referral traffic.
With the channel defined, view your traffic through the custom channel group and read the AI Assistants line. Track it over weeks and months to see whether AI-referred sessions are growing.
Segment further if useful — landing pages that AI sends traffic to often reveal which of your pages engines are citing. Pair the trend with your AEO work to see whether improvements move the number.
TrueCite's GA4 AI-traffic feature, available on starter and higher plans, surfaces AI-referred sessions without hand-building the channel group, and connects that traffic to your broader AI visibility. Either way, the goal is the same: make AI traffic visible.
Your AI channel group is only as good as the referrer list behind it. New AI engines appear, and existing ones change the domains they link from, so a static rule slowly drifts out of date.
Revisit your referrer report periodically and add any new AI hostnames sending you traffic. When you spot an unfamiliar source that looks like an assistant, fold it into your AI channel so the metric stays complete.
Also sanity-check the numbers. If AI traffic suddenly drops to zero, a domain may have changed or a rule may have broken rather than demand actually vanishing. Treating the channel group as something you maintain, not set once, keeps your AI traffic measurement honest over time.
AI referral traffic is measurable once you name the sources. List the AI referrer domains, build a custom GA4 channel group, define a rule that captures them, and watch the trend — a simple setup that turns invisible AI-driven visits into a metric you can track.
Using TrueCite? See the Trends docs →