How to Track ChatGPT & AI Traffic in GA4 [5-Min Setup]

Why Track AI Traffic?

Google Analytics doesn't automatically distinguish between AI referral traffic and regular website referrals. By default, visits from ChatGPT, Perplexity, and Claude get grouped into the generic "Referral" channel - or worse, show up as "Direct" traffic when referrer data isn't passed.

This matters because AI traffic behaves differently from other channels:

• Higher engagement - AI visitors spend 68% more time on websites than organic search visitors
• Better conversions - One study found AI traffic drove 12.1% more signups despite being only 0.5% of total visitors
• Different intent - Users coming from AI tools have already refined their query and are further along in their decision journey
• Rapid growth - AI referral traffic is growing 165x faster than organic search traffic

Without proper tracking, you can't measure this emerging channel, optimize your content for AI visibility, or report on its ROI to stakeholders. To complement your custom channel setup, check out GA4's new Generated Insights feature, which automatically surfaces your top data changes — including AI traffic anomalies.

AI Traffic Statistics 2026

Before diving into setup, let's understand the current landscape of AI referral traffic:

Market Share by Platform

• ChatGPT - 55-65% of all AI referral traffic (64.5% of Gen AI web traffic as of January 2026)
• Perplexity - 15-20% market share (170 million monthly visits)
• Google Gemini - 10-14% market share
• Microsoft Copilot - 6-9% market share (now passes referrer data to GA4)
• DeepSeek - 0.37% market share
• Claude - Growing share with 5% conversion rate on referred traffic

Growth Trajectory

• Overall AI traffic share - 1.08% of all web traffic, projected to reach 20-28% of total referral traffic by end of 2026
• ChatGPT referrals to news sites - 243.8 million visits in April 2025, continuing to accelerate into 2026
• Month-over-month growth - AI referral traffic growing roughly 1% month-over-month
• Comparison to search - Google still sends 300x more traffic, but AI referrals convert at 15.9% vs Google's 1.76%

ChatGPT Platform Stats

• Monthly visits - 6+ billion monthly visits
• Weekly active users - 800 million (doubled from 400M in February 2025)
• Global ranking - 6th most visited website worldwide
• New in 2026 - ChatGPT now appends utm_source=chatgpt.com to citation links, improving GA4 attribution

💡 Key Insight: While AI traffic is still small (1% of total), it's the fastest-growing referral channel. Companies tracking it now will have a significant data advantage.

Method 1: Create a Custom Channel Group (Recommended)

This is the best method because it makes AI traffic appear as its own channel in all standard acquisition reports, and it applies retroactively to historical data.

Step-by-Step Setup

1. Open GA4 Admin - Click the gear icon in the bottom-left corner of GA4
2. Navigate to Channel Groups - Under "Data Display," click "Channel Groups"
3. Create or Edit - Click "Create new channel group" or edit an existing custom group
4. Add New Channel - Click "Add new channel" and name it "AI Traffic" or "LLM Referrals"
5. Set the Condition:
   • Select "Source"
   • Choose "matches regex"
   • Paste the regex pattern (see below)
6. Reorder Channels - This is critical! Drag "AI Traffic" ABOVE "Referral" in the channel list
7. Save - Click save to apply the channel group

Why Order Matters

GA4 processes channel rules from top to bottom and assigns traffic to the first matching channel. If "Referral" comes before "AI Traffic," your ChatGPT visitors will be categorized as generic referrals before the AI rule is ever checked.

Basic Regex Pattern

chatgpt\.com|chat\.openai\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|copilot\.microsoft\.com

Method 2: GA4 Exploration Report

Use this method for ad-hoc analysis or when you want to dive deeper into AI traffic without creating permanent channel groups.

Step-by-Step Setup

1. Open Explorations - In GA4, click "Explore" in the left menu
2. Create Blank Report - Click "Blank" to start a new exploration
3. Name Your Report - Call it "AI Traffic Analysis" or similar
4. Add Dimensions:
   • Click the "+" next to Dimensions
   • Add "Session source" and "Session source/medium"
   • Add "Landing page" for content analysis
5. Add Metrics:
   • Sessions
   • Engaged sessions
   • Engagement rate
   • Average engagement time
   • Conversions (key events)
6. Apply Filter:
   • In the Settings panel, scroll to "Filters"
   • Add filter: Session source → matches regex
   • Paste your AI regex pattern
7. Build Visualization - Drag dimensions to Rows, metrics to Values

💡 Note: GA4 Exploration reports have date range limitations. For long-term trend analysis, use Looker Studio instead.

Method 3: Looker Studio Dashboard

Looker Studio (formerly Data Studio) is ideal for creating shareable AI traffic dashboards with no date limitations.

Setup Instructions

1. Create New Report - Go to lookerstudio.google.com and create a new report
2. Connect GA4 - Add your GA4 property as a data source
3. Create a Table - Add a table visualization
4. Configure Dimensions - Set "Session source" as the dimension
5. Add Metrics - Include Sessions, Users, Engagement Rate, Conversions
6. Add Filter:
   • Click "Add a filter"
   • Create new filter
   • Session source → Regex → Contains
   • Paste: chatgpt|perplexity|claude|gemini|copilot|anthropic|openai
7. Add Time Series - Create a line chart to visualize AI traffic growth over time

Dashboard Elements to Include

• Scorecard - Total AI sessions (with comparison to previous period)
• Pie Chart - Traffic breakdown by AI platform
• Time Series - AI traffic trend over time
• Table - Top landing pages from AI referrals
• Comparison - AI vs. Organic vs. Referral engagement metrics

Complete Regex Patterns for AI Platforms

Use these regex patterns to capture traffic from all major AI platforms. Update periodically as new platforms emerge.

Comprehensive Pattern (Recommended)

^(chatgpt\.com|chat\.openai\.com|perplexity\.ai|claude\.ai|anthropic\.com|gemini\.google\.com|bard\.google\.com|copilot\.microsoft\.com|copilot\.com|bing\.com/chat|you\.com|meta\.ai|poe\.com|pi\.ai|phind\.com|deepseek\.com|huggingface\.co|character\.ai|mistral\.ai|manus\.im|aisearchindex\.space)$

Simplified Pattern (Main Platforms Only)

chatgpt|perplexity|claude|gemini|copilot|anthropic|openai|deepseek|manus

Platform-Specific Patterns

• ChatGPT/OpenAI: chatgpt\.com|chat\.openai\.com|openai\.com
• Perplexity: perplexity\.ai
• Claude/Anthropic: claude\.ai|anthropic\.com
• Google AI: gemini\.google\.com|bard\.google\.com
• Microsoft: copilot\.microsoft\.com|copilot\.com|bing\.com/chat|edgeservices\.bing\.com
• AI Agents: manus\.im|aisearchindex\.space

💡 Pro Tip: GA4 regex is case-sensitive. Most AI referrers use lowercase, but test with your actual data to confirm.

Tracking Limitations & Workarounds

AI traffic tracking isn't perfect. Here are the main limitations and how to work around them:

Limitation 1: Missing Referrer Data

Problem: Free ChatGPT users and many AI tools don't pass referrer data. This traffic appears as "Direct" in GA4.

Workaround: Use UTM parameters when sharing links that might be referenced by AI tools:

https://yoursite.com/page?utm_source=chatgpt&utm_medium=ai_referral

Limitation 2: Copy-Paste Behavior

Problem: Users often copy URLs from AI responses and paste them into their browser, which breaks referrer tracking.

Workaround: Monitor your "Direct" traffic alongside AI referrals. Spikes in direct traffic to pages frequently cited by AI tools may indicate hidden AI referrals.

Limitation 3: New Platforms

Problem: New AI tools launch regularly, and your regex pattern won't capture them automatically.

Workaround: Periodically review your referral report for unfamiliar sources containing "ai," "chat," or "assistant" in the domain. Update your regex quarterly.

Limitation 4: ChatGPT UTM Inconsistency

Problem: ChatGPT sometimes adds utm_source=chatgpt.com but no utm_medium, causing "(not set)" in reports.

Workaround: In your custom channel group, add an additional condition for Source containing "chatgpt" to catch these variations.

Measuring AI Traffic Quality

Volume isn't everything. Here's how to measure whether AI traffic is actually valuable for your business:

Key Quality Metrics

• Engagement Rate - Compare AI traffic engagement rate vs. your site average
• Average Engagement Time - AI visitors typically spend 68% more time on site
• Pages per Session - Do AI visitors explore more content?
• Bounce Rate - Lower bounce typically indicates relevant traffic
• Conversion Rate - The ultimate quality metric

Creating a Comparison Segment

Create a GA4 audience called "AI Visitors" using the same regex pattern, then use GA4's comparison feature to benchmark AI traffic against organic search, paid, and referral channels. For more advanced segmentation, explore the new GA4 audience templates that let you build High-Value Purchaser and Disengaged Purchaser segments directly from pre-built templates.

For more on building comprehensive dashboards that include AI traffic, see our GA4 dashboard best practices guide. You can also use the GA4 attribution report to understand how AI-referred visitors convert compared to other channels.

Frequently Asked Questions

Can You Attribute Traffic from ChatGPT or Perplexity?

One of the most common questions in 2026 is whether you can reliably attribute website visits back to specific AI platforms like ChatGPT or Perplexity. The short answer: yes, but with caveats.

What's Changed in 2026

AI platforms are improving referral attribution. ChatGPT now appends utm_source=chatgpt.com to citation links, meaning GA4 can automatically attribute these visits without any custom setup. Microsoft Copilot also started passing referrer data to analytics platforms in late 2025, so copilot.com and copilot.microsoft.com now appear in your GA4 acquisition reports.

New AI Referral Sources to Watch

Beyond the major platforms, several emerging AI tools are now driving trackable referral traffic:

• Copilot.com — Microsoft's standalone AI assistant, now sending measurable referral visits separate from Bing Chat
• Manus.im — An autonomous AI agent platform that references and links to web content during task execution
• AIsearchindex.space — An AI search indexing service that surfaces and links to content across AI platforms

Attribution Gaps That Still Exist

Despite improvements, some AI traffic still falls through the cracks. Free-tier ChatGPT users on mobile often don't pass referrer data, causing visits to appear as "Direct." AI agents like Manus that browse autonomously may not always include referral headers. And when users copy a URL from an AI response and paste it into their browser, no referral data is passed at all.

To maximize attribution accuracy, combine three strategies: custom channel groups in GA4 (as described above), UTM parameters on links you control, and regular audits of your "Direct" traffic for spikes that correlate with AI platform mentions.

New AI Referral Sources to Track in 2026

The AI referral landscape has expanded significantly since this guide was first published. Beyond ChatGPT and Perplexity, several new platforms are now actively sending traffic to websites — and if you're not tracking them, you're flying blind on a growing slice of your audience.

Manus.im — The AI Agent That Browses for Users

Manus is an autonomous AI agent that browses the web on behalf of users, researching topics, comparing products, and compiling reports. Unlike chatbots that simply cite sources, Manus actively visits pages, reads content, and can drive genuine referral sessions. Its global ranking has climbed to the top 2,500 websites, with traffic growing 28% month-over-month in early 2026. In GA4, Manus traffic appears as manus.im / referral. Because Manus browses autonomously, it doesn't always pass referral headers consistently — so some visits may land in your "Direct" bucket. Add manus\\.im to your AI regex pattern to catch what it does send.

Copilot.com — Microsoft's AI Gateway

Microsoft Copilot now sends referrer data to GA4, appearing as copilot.microsoft.com / referral or copilot.com / referral. With 1.8 billion Copilot interactions per month, this is a significant and growing traffic source. Some early data suggests Copilot referrals convert at higher rates than other AI sources because users are often in a task-completion mindset. Make sure your regex pattern includes both copilot\\.microsoft\\.com and copilot\\.com to capture all variants.

Real Numbers: AI Referrals in Action

To illustrate how fast this channel is growing, here are real referral numbers from a marketing SaaS site in March 2026:

• Total AI referral sessions: 382 (23.1% of all traffic)
• ChatGPT: 289 sessions (75.7% of AI referrals)
• Claude.ai: 83 sessions (21.7% of AI referrals)
• Manus.im: 5 sessions (emerging but growing)
• Copilot.com: 4 sessions (newly trackable)
• OpenAI direct: 1 session

That's nearly 1 in 4 visits coming from AI platforms — a figure that would have been unthinkable even six months ago. If your custom channel group only covers ChatGPT and Perplexity, you're missing roughly 10% of your actual AI traffic. Update your regex pattern to include all current sources and audit it quarterly as new AI tools emerge.

Conclusion

AI referral traffic is the fastest-growing channel in digital marketing, yet most analytics setups completely miss it. By creating custom channel groups in GA4, you can finally see how much traffic ChatGPT, Perplexity, Claude, and other AI tools are sending to your website. AI is also transforming paid search—Google's AI Max for Search campaigns now use AI models to automatically optimize ad targeting and creative.

The setup takes about 10 minutes, applies retroactively to historical data, and gives you insights that your competitors likely don't have. Start tracking today, and you'll be ahead of the curve as AI-driven discovery becomes an increasingly important part of how people find content online.

Remember: AI visitors aren't just plentiful - they're valuable. With 68% higher engagement and 12% better conversion rates, this is traffic worth understanding and optimizing for.