Tracking AI Referral Traffic in GA4: Complete Setup Guide

Why Tracking AI Traffic Matters

If you have been following AI SEO best practices and optimizing your content for AI search engines, there is a good chance that ChatGPT, Perplexity, or Gemini is already sending visitors to your website. The question is: do you know about it?

Most GA4 setups lump AI traffic into the generic "Referral" channel -- or worse, into "Direct" or "Unassigned" when the referring AI platform does not pass a proper referrer header. This means you could be receiving hundreds of high-value visits from AI assistants without realizing it.

Here is why this matters:

AI traffic converts dramatically better. Research from Semrush shows that visitors arriving from AI referrals convert at 4.4x the rate of traditional organic search. When ChatGPT recommends your product or links to your guide, the user arrives with a built-in endorsement. They are not browsing a list of ten results -- they clicked a specific recommendation from a tool they trust.

AI referral traffic is growing at 326% year-over-year. ChatGPT alone drives 84.2% of all AI referral traffic, and that share is expanding as more users shift from Google to conversational AI for their research. If you are not measuring this channel, you are blind to the fastest-growing traffic source on the web.

You cannot optimize what you cannot measure. Tracking AI traffic separately lets you answer critical business questions: Which AI platforms cite us most? Which pages receive AI traffic? What is the revenue impact? Without dedicated tracking, these questions remain unanswered -- and your AI visibility strategy operates on guesswork.

AI Traffic Sources You Need to Track

Before configuring GA4, you need to know exactly which domains to look for. Here are the six primary AI traffic sources your tracking setup must cover:

| AI Platform | Referral Domain | Notes | |---|---|---| | ChatGPT | chatgpt.com | Largest source (84.2% of AI referral traffic). Automatically adds utm_source=chatgpt.com. | | Perplexity | perplexity.ai | Growing fast. Always includes source citations with direct links. | | Claude | claude.ai | Anthropic's assistant. Growing enterprise adoption, especially via Apple Safari integration. | | Google Gemini | gemini.google.com | Integrated into Google Search via AI Mode. May also appear as google.com with AI Overviews. | | Microsoft Copilot | copilot.microsoft.com | Built into Bing, Windows, and Office 365. May also appear as bing.com/chat. | | Grok | grok.com | xAI's assistant, integrated with X/Twitter. Growing US market share. |

A few important notes. Google Gemini traffic can be tricky to isolate because AI Mode responses appear within the Google Search interface -- the referrer may show as google.com rather than gemini.google.com. Similarly, Copilot traffic sometimes arrives under bing.com. Your channel grouping configuration (covered in Step 1) should account for these edge cases.

For a deeper look at how each platform sends traffic differently, see our guide on measuring AI referral traffic.

ChatGPT's Automatic UTM Behavior

Before diving into the setup steps, it is worth understanding a feature that ChatGPT provides out of the box.

When ChatGPT includes a hyperlink in its response and a user clicks it, ChatGPT automatically appends utm_source=chatgpt.com to the destination URL. This means that even without any GA4 configuration on your end, you will see ChatGPT traffic attributed correctly as a referral in your Source/Medium reports.

However, there are limitations:

• Only ChatGPT does this consistently. Perplexity, Claude, Gemini, and Grok do not reliably add UTM parameters to outbound links.
• The traffic still lands in the generic "Referral" channel. Without a custom channel grouping, ChatGPT traffic is mixed in with traffic from Reddit, Twitter, and every other referral source.
• You cannot customize the campaign or content parameters. ChatGPT sets utm_source but does not add utm_medium, utm_campaign, or utm_content -- limiting your analysis options.

This is why a custom channel grouping is essential: it pulls all AI sources into a dedicated channel, regardless of whether the AI platform adds its own UTM tags.

Step 1: Create a Custom "AI Search" Channel Grouping in GA4

A custom channel grouping tells GA4 to classify traffic from AI sources into its own dedicated channel -- separate from generic "Referral," "Organic Search," or "Direct." This is the single most important step for AI traffic tracking.

Follow these steps:

1. Open your GA4 property and navigate to Admin (gear icon in the bottom-left).
2. Under the Data display section, click Channel groups.
3. Click the Create new channel group button in the top-right corner.
4. Name your channel group something descriptive, such as "Custom Channel Group with AI Search."
5. You will see GA4's default channels listed (Organic Search, Direct, Referral, etc.). Click Add new channel.
6. Name the new channel "AI Search".
7. Set the matching condition to: Source matches regex:

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

8. Click Save channel, then click Save group.

Important notes:

• Custom channel groupings are not retroactive. They apply only to data collected after creation. Set this up as soon as possible.
• Place the "AI Search" channel above the default "Referral" channel in the priority list. GA4 evaluates channels from top to bottom and assigns traffic to the first match.
• You can add additional AI sources later by editing the regex pattern -- just append |newdomain\.com to the expression.
• To account for Copilot traffic that arrives as bing.com/chat, you may want to add a second condition: Session source contains bing.com AND Landing page contains /chat. However, this can be complex to maintain, so start with the basic regex and refine over time.

Step 2: Set Up UTM Parameters for AI Sources

While ChatGPT adds its own UTM tags automatically, you can improve tracking for all AI platforms by using UTM-tagged URLs in the content that AI models are likely to cite.

This works because when you publish content that AI models reference, the links in your content are what get passed to users. If those links include UTM parameters, GA4 will capture them regardless of whether the AI platform adds its own referrer data.

UTM structure for AI-focused content:

https://yoursite.com/page?utm_source=ai-search&utm_medium=referral&utm_campaign=ai-seo-2026&utm_content=guide-name

Recommended UTM values:

| Parameter | Value | Purpose | |---|---|---| | utm_source | ai-search | Identifies AI as the traffic source category | | utm_medium | referral | Keeps the traffic classified correctly | | utm_campaign | ai-seo-2026 | Groups all AI-related traffic for the year | | utm_content | {page-identifier} | Identifies which specific page was cited |

Where to place UTM-tagged links:

• In your FAQ schema answers -- when the answer includes a "learn more" link
• In your JSON-LD structured data -- the mainEntityOfPage and url fields
• In your llms.txt file -- the curated list of URLs you present to AI crawlers
• In external content -- guest posts, directory listings, and third-party mentions that AI models frequently reference

One caveat: do not overuse UTM tags on internal navigation links. They are specifically for URLs that external AI models will discover and pass to users.

Step 3: Build a Custom AI Traffic Report

Now that your channel grouping is in place, create a custom Exploration report that gives you a focused view of AI traffic performance.

Follow these steps:

1. In GA4, navigate to Explore in the left sidebar.
2. Click Blank to create a new exploration.
3. Name it "AI Referral Traffic Analysis."
4. In the Variables panel on the left, add the following dimensions by clicking the + button:
   • Session source
   • Session medium
   • Landing page + query string
   • Device category
   • Country
5. Add the following metrics:
   • Sessions
   • Engaged sessions
   • Engagement rate
   • Conversions (or Key events)
   • Bounce rate
   • Average session duration
   • Pages per session
6. In the Tab Settings panel:
   • Set Rows to "Session source"
   • Set Values to all the metrics you added
   • Add a Filter: Session source matches regex chatgpt\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|copilot\.microsoft\.com|grok\.com
7. Click Apply and your report will populate with AI-only traffic data.

Pro tip: Create a second tab in the same exploration with "Landing page" as the Row dimension. This shows you which specific pages on your site are being cited by AI models -- essential for understanding what content AI finds valuable.

For guidance on which pages to optimize based on this data, see our AI SEO checklist for 2026.

Step 4: Compare AI Traffic vs Organic vs Direct

One of the most valuable analyses you can run is a side-by-side comparison of AI traffic against your traditional channels. This reveals whether AI visitors behave differently -- and the data consistently shows they do.

To build the comparison report:

1. Go to Reports > Acquisition > Traffic acquisition.
2. Change the primary dimension dropdown from "Default channel group" to your custom channel group (created in Step 1).
3. You will now see "AI Search" as a distinct row alongside Organic Search, Direct, Referral, Social, and other channels.
4. Click the Comparison icon (two overlapping rectangles at the top) to add a comparison filter if you want to isolate specific date ranges.

What to look for:

| Metric | What it tells you | Typical AI traffic benchmark | |---|---|---| | Conversion rate | How AI visitors convert vs other channels | 4.4x higher than organic | | Bounce rate | Whether AI visitors engage or leave immediately | Usually 15-25% lower than organic | | Pages per session | Depth of engagement | Often 1.5-2x higher than organic | | Avg. session duration | Time spent on site | Typically 40-60% longer than organic |

If your AI traffic shows significantly worse performance than these benchmarks, it may indicate that the landing pages being cited by AI do not match user intent. Check which pages receive AI traffic (from your Exploration report) and ensure they deliver on the promise that the AI's citation implies.

Understanding these patterns is part of building a comprehensive AI visibility monitoring strategy.

Step 5: Key Metrics to Track

Not all metrics are equally important for AI traffic analysis. Focus on these core indicators:

Primary metrics

• AI Sessions -- Total visits from AI referral sources. This is your top-line growth indicator. Track week-over-week and month-over-month trends.
• AI Conversion Rate -- The percentage of AI visitors who complete a key event (purchase, signup, form submission). Compare this against organic and direct channels.
• AI Revenue / Value -- If you have ecommerce or goal values configured, track the total revenue attributable to AI traffic. This is your business case for continued AI SEO investment.

Engagement metrics

• Bounce Rate -- AI visitors should bounce less than organic visitors because they arrive with higher intent. A high bounce rate suggests a mismatch between AI citations and page content.
• Pages per Session -- Measures how deep AI visitors go into your site. Higher values indicate strong content relevance.
• Average Session Duration -- Time on site correlates with conversion probability. AI traffic typically shows longer sessions.

Growth and discovery metrics

• New vs Returning Visitors -- What percentage of AI traffic is new users? A high new-user rate means AI is driving brand discovery.
• Landing Page Distribution -- Which pages do AI platforms cite most? This reveals your strongest AI-visible content and helps you check if your site is visible in AI.
• AI Source Breakdown -- Track each AI platform individually. If 95% of your AI traffic comes from ChatGPT but zero from Perplexity, that is an optimization signal.

Step 6: Set Up Alerts for AI Traffic Spikes

AI traffic can spike suddenly when a model starts citing your content for a popular query. You want to know about this immediately -- both to capitalize on the opportunity and to ensure your site handles the traffic.

To create an AI traffic spike alert:

1. In GA4, go to Reports > Home (or the Reports snapshot).
2. Scroll down to the Insights section and click View all insights.
3. Click Create in the top-right corner to create a custom insight.
4. Configure the alert:
   • Name: "AI Traffic Spike"
   • Frequency: Daily
   • Segment: Create a segment where Session source matches regex chatgpt\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|copilot\.microsoft\.com|grok\.com
   • Metric: Sessions
   • Condition: "Is above" with a threshold of % increase that makes sense for your baseline (start with 50% increase over the previous period)
5. Toggle on email notifications and add the relevant team members.
6. Click Create.

Additional alerts to consider:

• AI conversion spike -- Triggered when AI traffic conversions exceed a threshold. Useful for ecommerce sites.
• New AI source detected -- Set a separate alert for each AI source individually. When a new platform starts sending traffic, you want to know which one.
• AI traffic drop -- Equally important. If ChatGPT stops citing you, a drop alert lets you investigate quickly. Check your robots.txt configuration first -- an accidental block is the most common cause.

Monthly Reporting Template

Consistent monthly reporting turns raw data into actionable insights. Use this template to structure your AI traffic reviews:

Section 1: Traffic Overview

| Metric | This Month | Last Month | Change | YoY Change | |---|---|---|---|---| | Total AI Sessions | --- | --- | ---% | ---% | | AI Share of Total Traffic | ---% | ---% | --- | --- | | New Users from AI | --- | --- | ---% | ---% |

Section 2: Conversion Performance

| Metric | AI Search | Organic Search | Direct | Paid | |---|---|---|---|---| | Sessions | --- | --- | --- | --- | | Conversion Rate | ---% | ---% | ---% | ---% | | Revenue / Value | --- | --- | --- | --- | | Avg. Session Duration | --- | --- | --- | --- |

Section 3: AI Source Breakdown

| AI Platform | Sessions | Conv. Rate | Top Landing Page | |---|---|---|---| | ChatGPT | --- | ---% | --- | | Perplexity | --- | ---% | --- | | Claude | --- | ---% | --- | | Gemini | --- | ---% | --- | | Copilot | --- | ---% | --- | | Grok | --- | ---% | --- |

Section 4: Top AI-Cited Pages

List the 10 pages receiving the most AI referral traffic. For each, note the primary AI source and whether the page has been specifically optimized for AI citation.

Section 5: Action Items

Based on the data, identify 3-5 specific actions for the next month:

• Which pages should be optimized for AI citation?
• Which AI platforms show untapped potential?
• Are there technical issues blocking AI crawlers?
• Should content be created to target topics where AI traffic is growing?

This template works well alongside your AI visibility monitoring workflow. Run the GA4 report on the same day you check your AI Visibility Score to get the complete picture.