How to Measure AI Referral Traffic in GA4

AI assistants increasingly send clickable citations, and those clicks land on your site as referral traffic. The problem is that GA4 doesn't group them under a tidy "AI" channel by default. Here's how to surface, segment, and report on AI referral traffic properly.

Why AI traffic hides in GA4

GA4's default channel groupings predate the AI-search boom. Visits from ChatGPT or Perplexity typically land in Referral (or sometimes Direct, when a referrer isn't passed), scattered among ordinary referrers. Unless you build a view for it, AI traffic stays invisible in your standard reports.

Step 1: Identify the referral domains

The main AI sources show up under these referrers:

• chatgpt.com and chat.openai.com (ChatGPT)
• perplexity.ai (Perplexity)
• (Gemini)

Start in Reports, open Acquisition, then Traffic acquisition, and set the primary dimension to Session source / medium. Search for these domains to confirm which are already sending traffic.

Step 2: Build an exploration to isolate them

The cleanest approach is a custom exploration:

1. Open Explore and create a Free-form exploration.
2. Add Session source as a dimension and Sessions, Engaged sessions, and Conversions as metrics.
3. Add a filter where Session source matches a regex like .

This gives you a single table of AI-driven sessions and how they behave.

Step 3: Create a custom channel group (optional but powerful)

If you want AI traffic in your standard reports, build a custom channel group in Admin:

• Create a new channel called AI Assistants.
• Define the rule: Source matches regex chatgpt|openai|perplexity|gemini|copilot|claude.
• Order it above Referral so those sessions get claimed first.

Now AI traffic appears as its own channel going forward.

Step 4: Watch for the Direct-traffic gap

Some assistants strip the referrer, so a share of AI clicks land in Direct. You won't capture those with source rules alone. Two partial fixes:

• Encourage citation formats that preserve referrers where you control them.
• Use UTM parameters on links you place in AI-readable content and documentation.

Accept that AI attribution will be directional, not perfect.

Step 5: Track the metrics that matter

Raw sessions are only the start. For AI referrals, watch:

• Engagement rate and average engagement time, since AI visitors often arrive pre-qualified.
• Conversions, to prove the channel's value.
• Landing pages, to learn which content gets cited and clicked.
• Trend over time, since AI referral share is climbing month over month for most sites.

Connect clicks back to citations

GA4 tells you when AI traffic arrives, but not why. To connect the dots, pair it with citation data: use SEO Roger's AI Citation Tracker to see which queries cite you, then match those against the landing pages earning clicks in GA4. When citations rise and referral sessions follow, you've closed the loop on impact.

Takeaway

AI referral traffic is real and growing, but GA4 won't hand it to you neatly. Identify the referrer domains, build an exploration or custom channel group to isolate them, accept some Direct-traffic leakage, and focus on engagement and conversions. Then tie those clicks back to the citations that produced them so you can invest where AI search actually pays off.