Key takeaways
- ChatGPT adds utm_source=chatgpt.com to links in its answers; other assistants usually pass a referrer domain.
- A custom channel group in GA4 with an “AI Assistants” channel placed above “Referral” makes this traffic visible.
- Capture the same source on forms so AI-sourced leads reach your CRM and pipeline reports.
- Clicks from Google AI Overviews and AI Mode are reported as Google organic and cannot be separated in GA4.
- Some AI influence never leaves a referrer — add “AI assistant” to self-reported attribution.
Why AI traffic is hard to see
When someone clicks a cited link in an AI answer, GA4 attributes the session by its source. Without a rule for AI sources, those sessions fall into “Referral”, mixed with every other website — or into “Direct” when the app strips the referrer. A dedicated channel fixes the first problem; form attribution and self-reported source help with the second.
Which signals each assistant sends
| Assistant | Typical source in GA4 | Notes |
|---|---|---|
| ChatGPT | chatgpt.com | OpenAI adds utm_source=chatgpt.com to links |
| Perplexity | perplexity.ai | Referrer from web; apps may differ |
| Gemini | gemini.google.com | Referrer from web app |
| Microsoft Copilot | copilot.microsoft.com | Some clicks arrive as bing.com |
| Claude | claude.ai | Referrer from web app |
| Google AI Overviews / AI Mode | google / organic | Not separable in GA4 |
Step 1 — Create an “AI Assistants” channel group
- In GA4, go to Admin → Data display → Channel groups and create a new group by copying the default one.
- Add a channel named AI Assistants with the condition Session source → matches regex:
.*(chatgpt\.com|openai\.com|perplexity\.ai|gemini\.google\.com|copilot\.microsoft\.com|claude\.ai|deepseek\.com|meta\.ai|grok\.com|you\.com|kimi\.com|doubao\.com).*- Move the channel above “Referral”. GA4 assigns a session to the first channel whose rule matches, so order matters.
- Save and use the new group in Traffic acquisition and Explorations. Custom channel groups also apply to historical data.
Step 2 — Carry the source into your CRM
Analytics shows sessions; the CRM shows revenue. Add hidden fields to every form for utm_source, utm_medium, utm_campaign, referrer and landing page, and map them to lead-source properties in HubSpot or Salesforce. Flag a lead as AI-sourced when the source matches the same regex. (This website does exactly that, after visitor consent.)
Step 3 — Build the reports that matter
- Landing pages by AI source — which pages earn citations that people actually click.
- Engagement and conversion rate of AI sessions versus organic and paid.
- AI-sourced leads → SQLs → pipeline in the CRM, monthly.
Step 4 — Watch the crawlers, too
Referral sessions are only half the story. Server logs or your CDN’s bot analytics show how often OAI-SearchBot, PerplexityBot, Claude-SearchBot and others fetch your pages. A page that is never crawled cannot be cited.
The blind spots
- Google AI Overviews and AI Mode clicks appear as Google organic.
- Mobile apps and copy-pasted URLs often arrive without a referrer.
- Many buyers read an AI answer, then search your brand name later — the visit shows as branded search or direct.
Add “ChatGPT or another AI assistant” to your “How did you hear about us?” field and watch branded-search trends alongside AI visibility. Our GEO Measurement Framework combines all of these signals.