Your website may already receive visitors from ChatGPT or Perplexity, but your usual SEO report can miss them. To track AI search referrals in GA4, we start with session-level source data, then connect those visits to leads and sales.
The setup is straightforward, but the numbers need context: GA4 only identifies AI referrals when it receives usable attribution information. First, let’s separate the visits you can measure from the discovery activity your reports can’t see.
What GA4 Can and Can’t Track From AI Search
GA4 can identify an AI platform when a visitor follows a link that passes referrer information. That visit may appear with a source such as chatgpt.com and a medium such as referral.
However, source formats aren’t consistent across every platform, browser, and app. Missing referrers, copied links, and privacy settings can leave an AI-influenced visit without an identifiable AI source.
Google’s explanation of campaigns and traffic sources describes how Analytics processes these signals. We use that distinction to keep reporting realistic.

Clicks from Google’s AI Overviews and AI Mode can arrive as google / organic. That source alone doesn’t separate AI-feature clicks from other Google Search clicks.
GA4 also doesn’t reveal the prompt that produced a referral, every citation your business received, or answers viewed without a click.
A dedicated AI channel organizes identifiable visits. It can’t recover AI discovery that arrived without a usable source.
We treat observable AI referrals as one part of SEO performance, not a complete measure of AI visibility.
Find AI Search Referrals in GA4 Traffic Acquisition
Start with Traffic acquisition rather than User acquisition. We want to understand the source attributed to a session and the actions recorded during that visit.
Choose Session Source/Medium
Open Reports, find Acquisition, and select Traffic acquisition. Depending on your report collection, Acquisition may appear under Life cycle or a business-objective grouping.
Change the table’s primary dimension to Session source/medium. Set a date range with enough activity to review, such as the last 28 days.
Google’s traffic-source dimension guidance explains the different scopes. First user source describes initial acquisition; session dimensions are the useful starting point for this workflow.

Search for Recognizable AI Sources
Search the table for known AI domains. These are useful starting points, not a permanent list.
| AI platform | Source to check |
|---|---|
| ChatGPT | chatgpt.com |
| Perplexity | perplexity.ai |
| Claude | claude.ai |
| Google Gemini | gemini.google.com |
| Microsoft Copilot | copilot.microsoft.com |
Review the actual source/medium combinations before building filters. Don’t require the medium to equal referral: identifiable AI sources can appear with other values.
Record the source spelling, sessions, engagement rate, and relevant key events. Then check whether those visits reached useful pages.
Zero results mean GA4 has no matching recorded sessions for that period. They don’t prove your business never appeared in an AI answer.
Create a Custom Channel for AI Referrals
Searching individual domains works for an initial check. A custom channel group makes ongoing comparisons easier.
We recommend keeping the default channel group available while adding a separate view for identifiable AI sources.
Add Source-Based Channel Rules
Open Admin, then Data display, then Channel groups. Choose Create new channel group and give the group a clear name, such as “SEO Acquisition.”
Add a channel named “AI Assistants.” Configure its Source condition using the domains you’ve verified in your reports.
For exact domain matching, a starting regular expression is:
^(chatgpt\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|copilot\.microsoft\.com)$
This matches those exact source values. Add other verified domains or subdomains as separate alternatives when they appear.
Google’s custom channel group instructions explain how to create channels and arrange their rules.
Put the AI Channel Before Broader Matches
Move “AI Assistants” above broader channels that could capture the same sessions. Channel rules are evaluated in order, so placement matters.
We don’t add a required referral medium condition unless there’s a reporting reason. A source-first rule keeps recognizable AI traffic together despite inconsistent mediums.
Save the group, then select its session-scoped dimension in Traffic acquisition. Compare its sessions with the source rows you checked earlier.
Custom channel rules can classify historical source data already available in GA4. They don’t change the original source values or reveal missing referrers.
Revisit the rules monthly. If a new AI source appears, verify it before adding it to the channel.
Build a Landing-Page Report in Explorations
Next, connect AI referrals to the pages people enter. This helps you see which content attracts visits and what happens afterward.
Open Explore and create a free-form exploration. Import Session source, Session medium, and Landing page + query string as dimensions.
Add Sessions, Engaged sessions, Engagement rate, Key events, and Total revenue as metrics where relevant. Google’s dimensions and metrics reference provides definitions for these reporting fields.
Then assemble the report:
- Place Landing page + query string in Rows and Session source in Columns.
- Add your selected metrics to Values.
- Filter Session source using the same verified AI-domain expression.
- Save the exploration with a clear name and review period.
Keep the date range consistent with Traffic acquisition. Also use session-scoped source dimensions throughout, rather than switching to first-user acquisition halfway through the comparison.
Don’t expect UTM parameters to appear in Landing page + query string. Use the appropriate campaign dimensions when reviewing tagged visits.
We look for pages that attract engaged sessions and useful outcomes. A heavily visited article may need a clearer next step, while a lower-volume service page may produce stronger leads.
Measure Leads and Revenue Alongside AI Traffic
Traffic counts help you spot activity. Business outcomes help you decide where to invest.
We recommend measuring the same completed actions across every channel. Then filter reporting to compare AI referrals with organic search, paid traffic, and other sources.
Track Completed Actions as Key Events
Use reliable events for successful form submissions, confirmed appointments, and purchases. Mark the actions that matter as key events in GA4.
A generate_lead event should reflect a completed lead action, not a click on the submit button. Enhanced measurement can provide useful behavior signals, but form events still need validation.
Keep phone-link clicks separate from completed calls. A phone_link_click event shows intent; it doesn’t prove someone reached your team.
Call-tracking platforms such as CallRail can provide additional call information when configured to retain attribution. For sales reporting, connect those outcomes with your CRM. Our guide to SEO lead attribution explains that connection.
Compare Rates and Lead Quality
Review relevant key events, session key event rate, and revenue beside sessions. Choose the key event you’re evaluating instead of blending unrelated actions.
Small samples need patience. One sale from a handful of AI sessions doesn’t establish a reliable channel conversion rate.
For lead-generation businesses, ask whether inquiries were qualified, appointments happened, and jobs were booked. An estimated lead value is useful for planning, but keep it separate from actual revenue.
Then review Google visibility alongside post-click behavior. Our Search Console Insights guide helps connect search trends with GA4 outcomes without treating the two tools as interchangeable.
Validate Tracking Before Trusting the Report
A useful report depends on a reliable implementation. We check collection first, then attribution.
Test Tags and Lead Events
Confirm the GA4 tag runs on relevant pages. Use Realtime for a quick collection check, then DebugView or Google Tag Manager preview for event testing.
Submit a test form and confirm that its success event appears once. Duplicate tags or overlapping triggers can inflate results.
Record events for all visitors, then segment by source in reporting. Restricting a lead-event tag to organic visitors prevents useful comparisons with AI referrals.
Also check consent behavior. Consent choices and blockers can limit what GA4 observes; they aren’t gaps that a channel rule can repair.
Check Referrers and Redirects
Don’t add AI platforms to the unwanted-referrals list. Suppressing their referrers can hide the source information you’re trying to measure.
Review redirects and cross-domain journeys when source data looks wrong. A booking system or payment service can interrupt attribution if the implementation isn’t configured correctly.
Test an actual link from an AI platform when available, ideally in a fresh browser session. Capture the landing page, then check the processed session data later.
A single test confirms that one click path works. Different apps and browsers can behave differently.
Key Takeaways for AI Referral Reporting
We keep the reporting routine focused on three priorities:
- Use session-scoped source dimensions to identify recorded AI visits without requiring a referral medium.
- Group verified AI sources, then compare landing pages, engagement, and meaningful key events.
- Review qualified leads and revenue alongside traffic, while keeping attribution limits visible.
Maintain the domain rules and review performance monthly. That gives you a repeatable process without treating every short-term change as an SEO success.
FAQ About Tracking AI Search Traffic
Does GA4 Show Every Visit Influenced by AI?
No. GA4 needs usable attribution signals to identify a platform as the source.
Someone may discover your business in an AI answer, then search your name or type your address later. That journey doesn’t automatically appear as an AI referral.
We report identifiable AI sessions separately and use other evidence, such as customer feedback and visibility monitoring, for the broader picture. Neither one should be presented as a complete record of discovery.
Should AI Referrals Count as Organic Search?
We keep identifiable AI-assistant referrals in a separate custom channel for clearer comparisons.
GA4’s default classification depends on the source and medium it receives. Your custom channel is a reporting category, not proof of the visitor’s full journey.
Google AI-feature clicks that appear as google / organic need separate context. Don’t move all Google organic traffic into an AI channel. That would mix ordinary search clicks with AI-assisted search activity.
Turn AI Referral Data Into Better SEO Decisions
Tracking AI search referrals in GA4 gives you a practical view of identifiable visits and their outcomes. The strongest report connects source, landing page, and business results.
Start with the source data your property records. Then maintain the channel rules and validate the actions you measure.
Your next SEO decision should follow the pages that bring qualified demand, with the limits of attribution kept clear.





