How to Track AI Traffic in GA4: Referrers, UTMs, and Custom Channels

AI traffic attribution is messy because AI interfaces were not designed as clean marketing channels. Some pass referrers, some do not, and some influence users without any click at all. GA4 can still capture part of the picture if the setup is explicit.

Why GA4 Misses AI Traffic

GA4 relies on referrer, UTM parameters, and campaign rules. AI products often break that chain. A link opened from a mobile app can lose the referrer. A user can copy a URL from an AI answer. A user can read the recommendation, close the AI tool, and later search for the brand.

That is why AI traffic is usually split across Direct, Referral, Organic Search, and branded organic.

Referrer Domains to Track

Use this starting list and update it as products change:

chatgpt.com
openai.com
perplexity.ai
claude.ai
gemini.google.com
copilot.microsoft.com
bing.com
you.com
phind.com
duckduckgo.com
ya.ru
yandex.ru
grok.com
x.ai

Perplexity and You.com tend to pass referrers more reliably. ChatGPT and Copilot can be less consistent, especially from apps.

Create a Custom Channel Group

In GA4, go to Admin, Data Settings, Channel Groups. Create a new group and add a channel named AI Search.

Use a source regex such as:

(chatgpt|openai|perplexity|claude|gemini|copilot|bing|you\.com|phind|duckduckgo|ya\.ru|yandex|grok|x\.ai)

Place the AI Search channel above Organic Search and Referral. Channel order matters: the first matching rule wins.

Add UTM Parameters Where You Can

If you publish links that AI systems may quote, use explicit UTMs on controlled deep links:

utm_source=chatgpt
utm_medium=ai-referral
utm_campaign=ai_citation

For broader evergreen content, use source-specific values when possible: perplexity, copilot, gemini, or ai_search. UTMs cannot solve every attribution issue, but they reduce dependency on referrer behavior.

Build an Exploration Report

Create a Free Form Exploration with these dimensions:

• Session source / medium
• Landing page
• Page referrer
• Default channel group
• Device category

Use metrics such as sessions, engaged sessions, engagement rate, average engagement time, key events, and revenue. Add a comparison segment for AI Search and compare it with Organic Search.

Use BigQuery for Deeper Analysis

If volume is high, export GA4 to BigQuery. This allows cohort analysis, conversion-path analysis, and correlation between AI referrer sessions and later branded sessions. BigQuery is especially useful when standard GA4 reports hide detail through sampling or aggregation.

What GA4 Cannot Show

GA4 cannot tell you whether ChatGPT mentioned your brand without a click. It cannot show whether Perplexity cited a competitor instead of you. It cannot reliably attribute a user who saw an AI recommendation and returned later through branded search.

That upstream layer requires AI visibility monitoring. GEO Scout tracks the prompts, providers, mentions, positions, sentiment, and cited sources so analytics teams can interpret GA4 traffic in context.

Setup Checklist

• Create an AI Search channel group.
• Add source regex for known AI referrers.
• Place AI Search above Organic and Referral.
• Use utm_medium=ai-referral for controlled links.
• Create an Exploration report for AI sessions.
• Compare AI traffic with organic and direct traffic.
• Correlate GA4 trends with GEO Scout Mention Rate.

Practical Priority

Start with channel grouping and an Exploration report. Then add UTMs to pages AI is likely to cite: guides, comparison pages, pricing pages, research reports, and templates. Finally, connect traffic data with AI visibility data so clicks are not interpreted without the answer context that created them.