How do you track AI search traffic in Google Analytics 4?

To track AI search traffic in Google Analytics 4, create a custom channel group with an “AI Search” channel whose condition matches session sources such as chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com and claude.ai using a regular expression, and order it above Referral. Then report sessions, engagement and conversions for that channel, knowing the number is a floor.

How AI traffic reaches GA4

flowchart TD
    A[User clicks a link in an AI answer] --> B{Referrer or UTM passed?}
    B -->|Yes: chatgpt.com, perplexity.ai, etc.| C[Session source = AI domain]
    B -->|No: app, privacy settings, copy-paste| D[Session lands in Direct]
    C --> E{Channel group rule matches?}
    E -->|Yes| F[AI Search channel]
    E -->|No| G[Default Referral channel]
    H[Click from Google AI Overviews or AI Mode] --> I[google / organic, not separable]

By default, GA4 lumps AI referrals into the Referral channel. A custom channel group pulls them out.

Step 1: Check which AI sources you already get

1. Go to Reports > Acquisition > Traffic acquisition.
2. Change the primary dimension to Session source.
3. Search the table for “chatgpt”, “perplexity”, “gemini”, “copilot”, “claude”, “deepseek” and “grok”.

Note the exact source values you see. Common ones:

Assistant	Typical session source
ChatGPT	chatgpt.com (older: chat.openai.com)
Perplexity	perplexity.ai, www.perplexity.ai
Gemini	gemini.google.com
Microsoft Copilot	copilot.microsoft.com
Claude	claude.ai
DeepSeek, Grok	chat.deepseek.com, grok.com

Step 2: Create a custom channel group

1. Go to Admin > Data display > Channel groups.
2. Click Create new channel group. GA4 copies the default channels, which you should keep.
3. Name it, for example, “Default + AI Search”.
4. Click Add new channel and name it AI Search.
5. Add a condition: Session source > matches regex, and paste:

.*(chatgpt|openai|perplexity|gemini\.google|copilot\.microsoft|claude\.ai|deepseek|grok).*

6. Save the channel.
7. Reorder so AI Search sits above Referral (and above Organic Search). GA4 assigns each session to the first channel whose rule it matches, so order matters.
8. Save the channel group.

Adjust the regex to the sources you actually saw in Step 1, and review it quarterly as new assistants appear. Google’s help center explains custom channel groups in more detail.

Step 3: Report on it

• In Traffic acquisition, switch the dimension to your new channel group (for example “Session Default + AI Search”).
• Add key events (demo requests, signups, trials) as columns to compare conversion rates with Organic Search and Paid Search.
• Build an Exploration with Session source, Landing page and key events to see which pages AI assistants send traffic to.
• Add Landing page as a secondary dimension: AI referrals often land on pricing, comparison and docs pages, which tells you what engines cite.

Step 4: Connect it to pipeline

GA4 shows sessions and key events, but B2B revenue lives in the CRM.

• Capture first-touch source and the AI channel into hidden form fields and pass them to your CRM.
• Add “AI assistant (ChatGPT, Perplexity, Gemini, Copilot, etc.)” to “How did you hear about us?”.
• Report AI-sourced pipeline next to GA4 numbers. See how to calculate the ROI of GEO.

Optional refinements

Once the basic channel works, a few additions make the data more useful:

• Per-assistant breakdown. Keep Session source as a secondary dimension inside the AI Search channel so you can compare ChatGPT, Perplexity, Gemini and Copilot separately.
• Landing page groups. Group landing pages by type (pricing, comparison, docs, blog) to see which content AI assistants send people to and which converts.
• Engagement comparison. Compare engagement rate and key event rate against Organic Search. AI-referred visitors often arrive further along in their research, so differences here are worth reporting.
• Looker Studio dashboard. Pull the channel group into a simple dashboard alongside your AI visibility metrics so leadership sees traffic and visibility side by side.
• Annotations. Note dates of major content changes or crawler fixes so you can relate them to traffic shifts.

The limits you should report

Be explicit about what this setup cannot see:

1. Missing referrers. Clicks from desktop and mobile assistant apps, some browsers and privacy settings often arrive without a referrer and land in Direct. ChatGPT’s practice of adding utm_source=chatgpt.com to many links helps, but not every engine does this.
2. Google AI Overviews and AI Mode. These clicks are reported as google / organic. There is no reliable way to separate them in GA4.
3. No-click influence. Buyers who read an answer and later search your brand or type your URL appear as branded organic or Direct.
4. Copy-paste. Users who copy a URL from an answer arrive as Direct.
5. Sources change. New assistants and new domains appear; an outdated regex silently undercounts.

Because of these gaps, pair GA4 with visibility metrics: mention rate, citation rate and share of voice inside AI answers.

Setup checklist

• Session sources from AI assistants identified.
• Custom channel group created with an AI Search channel.
• Regex covers all observed AI sources.
• AI Search ordered above Referral and Organic Search.
• Key events added to the AI channel report.
• Source passed to CRM; “AI assistant” added to attribution forms.
• Regex reviewed quarterly.
• Limits documented in every report.

Where this fits in GEO reporting

AI referral sessions and conversions are the business half of GEO reporting. The visibility half, how often you are named and cited, needs prompt monitoring. Both are covered in the most important KPIs for GEO and on the GEO metrics and KPIs page. GrackerAI includes analytics that connect AI citations with the traffic they drive, if you want both views in one place.