How to Track AI Search Traffic in Google Analytics 4: Referral Sources, Sessions, and Attribution

AI search platforms are sending real, measurable traffic to websites. The challenge is that most GA4 setups are not configured to capture and report this traffic accurately: some of it lands in referral channels, some blends into organic Google traffic, and some disappears into direct/none. If you are trying to demonstrate the business value of AI search optimization, you need accurate AI traffic attribution data.

This guide walks through exactly how to configure GA4 to capture, segment, and report AI search traffic from the major AI platforms, including the custom channel group setup that makes AI traffic a first-class reporting dimension.

What you will learn:

• Which AI platforms send referral traffic and what their source domains are in GA4
• How to create a custom AI channel group in GA4 that consolidates AI traffic sources
• How to approximate Google AI Overview traffic despite Google's unified organic attribution
• How to build a simple AI traffic dashboard in GA4 for regular monitoring
• How to connect AI traffic data to conversion and pipeline attribution

AI Referral Sources in GA4: Where Each Platform Appears

Different AI platforms appear differently in GA4's Traffic Acquisition report. Understanding each one is the first step to accurate tracking.

ChatGPT (OpenAI)

ChatGPT traffic arrives from two source domains:

• chat.openai.com: standard ChatGPT conversation interface
• chatgpt.com: ChatGPT's dedicated domain (launched 2024)

Both appear as Referral traffic in GA4's default channel grouping. Search for these domains in the Traffic Acquisition report filtered by Session source/medium.

Perplexity AI

Perplexity traffic arrives from perplexity.ai and typically appears as a referral source. Perplexity provides inline citations that users click, making it one of the most cleanly trackable AI traffic sources. It often appears as one of the top referral sources for sites that have optimized for Perplexity citation.

For context on how Perplexity's citation model works, see How to Rank in Perplexity AI.

Claude (Anthropic)

Claude traffic is less consistently tracked because Claude's interface does not always generate a referral header when users copy and navigate to URLs. Traffic from Claude typically arrives from claude.ai when it does send a referral. However, a portion of Claude-attributed visits arrive as direct traffic. This "dark AI traffic" problem is most pronounced with Claude.

Google Gemini

Gemini traffic arrives from gemini.google.com. It appears as a Referral source in GA4. Note that this is distinct from Google AI Overviews: Gemini is a separate product, not the same as the AI Overviews that appear in Google Search.

Microsoft Copilot / Bing Chat

Copilot traffic arrives from:

• copilot.microsoft.com: standalone Copilot product
• bing.com: may appear as organic if users access Copilot through Bing search

The Bing-originating Copilot sessions may blend with general Bing organic traffic in some configurations.

Google AI Overviews

This is the most significant tracking gap. Google AI Overview traffic arrives through the same google.com / organic channel as all other Google organic traffic. There is currently no GA4 parameter that isolates AI Overview clicks from standard organic clicks. The workaround, discussed below, uses Search Console data as a proxy.

Setting Up a Custom AI Channel Group in GA4

Custom channel groups in GA4 allow you to define your own traffic classification rules, including an "AI Search" channel that consolidates all identified AI referral sources.

Step 1: In GA4, navigate to Admin > Data display > Channel groups.

Step 2: Create a new channel group (or edit the Default channel group).

Step 3: Add a new channel called "AI Search" with the following conditions (using OR logic):

Step 4: Position the "AI Search" channel above "Organic Search" and "Referral" in the channel priority order so AI sources are correctly attributed before falling into generic categories.

Step 5: Apply and allow 24 to 48 hours for the new grouping to populate in historical comparison views.

Approximating Google AI Overview Traffic

Since Google AI Overview traffic is not separately tagged in GA4, the best approximation uses Search Console data correlated with GA4 session data.

Method 1: Search Console CTR analysis for AI Overview candidate pages. Pages that appear in Google AI Overviews often show an unusual pattern in Search Console: impressions remain stable or grow while CTR increases beyond what the organic position change would predict. This over-performance relative to position is a directional indicator of AI Overview citation benefit.

Filter your Search Console Performance report to specific pages you know are appearing in AI Overviews. Track their impression-to-click ratio month-over-month. Improving CTR while position stays flat is the clearest Search Console signal of AI Overview citation value.

Method 2: Segment branded vs non-branded traffic patterns. AI Overview citations frequently increase branded traffic (users who encounter your brand in an AI Overview and then search specifically for your brand). Track branded query traffic in Search Console alongside total AI referral traffic in GA4. Correlated growth in both dimensions suggests AI Overview citations are creating downstream branded search demand.

Building an AI Traffic Dashboard in GA4

A simple AI traffic dashboard requires three report components:

Component 1: AI Search Traffic Trend

• Report: Traffic Acquisition
• Filter: Channel group = "AI Search"
• Dimensions: Date, Source
• Metrics: Sessions, Engaged sessions, Average engagement time, Conversions
• Date comparison: Current 30 days vs prior 30 days

Component 2: AI Landing Page Performance

• Report: Pages and Screens (filtered to Landing Page)
• Filter: Channel group = "AI Search"
• Dimensions: Landing page, Source
• Metrics: Sessions, Bounce rate, Conversions
• This shows which pages are receiving AI-sourced traffic and how they are converting

Component 3: AI Traffic Conversion Comparison

• Report: Conversions, filtered to compare "AI Search" vs "Organic Search" channel groups
• Metrics: Conversions, Conversion rate, Revenue (if applicable)
• Shows whether AI-sourced visitors convert at similar rates to Google organic visitors

For teams using Looker Studio for client reporting, these three components can be assembled into a single AI search performance dashboard that updates automatically from your GA4 connection.

Connecting AI Traffic to Pipeline Attribution

For B2B companies where website visits are not the final conversion metric, connecting AI referral traffic to pipeline requires CRM integration.

Step 1: Ensure your CRM capture (HubSpot, Salesforce, etc.) records the UTM source or referral source from the session where a lead first submitted a form.

Step 2: Tag or segment leads in your CRM where the first-touch source is an AI platform (perplexity.ai, chat.openai.com, etc.).

Step 3: Track these leads through your pipeline stages to measure AI-sourced lead quality, sales cycle length, and deal close rate compared to other acquisition channels.

Step 4: Use this data to calculate the AEO ROI formula covered in the AEO ROI measurement guide.

For a comprehensive framework for converting AI traffic data into business impact reporting, the AEO KPIs guide covers the metrics that translate AI search performance into executive-facing language. The AEO Monitoring and Tracking Guide covers citation probe monitoring that complements the GA4 traffic attribution data with AI-side measurement.