How to see ChatGPT and Perplexity traffic in GA4 (2-minute setup)

Visitors arriving from ChatGPT, Perplexity, Gemini and Copilot don’t announce themselves. GA4 files them under generic referrals — chatgpt.com here, perplexity.ai there — scattered across your reports where nobody adds them up. The fix takes two minutes.

The setup

GA4 lets you group traffic with custom channel groups using regex matching on the referral source. Create a channel called AI Assistants, match it on the referral domains of every AI surface (chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com, claude.ai and friends), and every past-and-future session from those sources rolls into one line you can trend, compare, and report.

The exact regex — maintained as new AI domains appear — plus click-by-click GA4 instructions are in our free GA4 AI traffic tool. Copy, paste, done.

What to expect when the data lands

• Volume will look small. Conversion won’t. Across reports we see, AI-referred visitors convert dramatically better than average — they arrive pre-sold by a recommendation, not browsing.
• It’s undercounted. Many AI answers get acted on without a click (the user just searches your brand next — that shows up as organic or direct). Treat the AI channel line as a floor, not a ceiling.
• The trend matters more than the total. This channel compounds. The month-over-month slope is the number to watch.

The other half of the picture

GA4 shows you the clicks you got. It can’t show you the recommendations you didn’t get — every answer where a competitor was named and you weren’t. That’s the half Asheard measures: which questions you appear for, across 9 AI engines, over time.