How to Track AI Traffic in GA4 (2026): Step-by-Step Guide

Key Takeaways

• GA4 does not label AI traffic by default:Visits from ChatGPT, Gemini, Perplexity, Copilot, and Claude are often categorized under Referral, Direct, or Unassigned, so you need custom settings to track AI traffic accurately.
• Track AI traffic in GA4 in three ways:Use Traffic acquisition to quickly establish a baseline, use Exploration with regex filters for ongoing analysis, and apply UTM parameters to links you control to improve attribution.
• Create a custom AI channel group:Use Session source regex to add an AI traffic channel group, place it above Referral, and save a report that shows only AI traffic for easier team reporting.
• Look beyond traffic volume:Review engagement and conversion performance, update AI source rules regularly, and use WorkDuo to track AI visibility, sentiment, and competitor movement end to end.

GA4 does not treat “AI traffic” as a standalone category by default. Visits from ChatGPT, Gemini, or other AI assistants are often mixed into Referral, Direct, or “Unassigned,” making AI search nearly invisible in your reports.

This creates a real measurement problem. You cannot optimize what you cannot see, and you cannot prove the value of AI search investment if the traffic is buried in messy reports.

Adobe found that traffic from AI referrals grew in the U.S. from July 2024 to February 2025 bymore than 10x, making proper GA4 setup a requirement, not a nice-to-have.

How to Track AI Traffic in GA4 (Step-by-Step Guide)

Here are the exact steps to follow to identify and report on AI traffic in GA4.

Step-by-Step Setup (Start Here)

Step 1: Identify AI referral sources

The first step is to identify common referral domains associated with AI platforms. Here are a few common examples:

• chat.openai.com (ChatGPT)
• perplexity.ai (Perplexity AI)
• gemini.google.com (Google Gemini)
• copilot.microsoft.com (Microsoft Copilot)
• claude.ai (Anthropic Claude)

💡 Pro tip:You can usually find these domains by reviewing your existing GA4 referral traffic reports. Look for unusual or new referral sources that do not fit traditional categories.

Step 2: Confirm GA4 can “see” AI referral traffic (quick test)

Before building reports, run a quick test to confirm GA4 is receiving data from AI referral sources.

1. In GA4, go to Reports → Realtime.
2. Click a link to your website from an AI tool, such as ChatGPT, then check Traffic source in Realtime to confirm whether referral source information is being passed. It often appears as a referral domain, but not always.

📌 Note:Some AI apps and AI browsers may remove or modify referral source data, so some AI-driven visits may be classified as Direct unless you tag the links.

Step 3: Remove noise to prevent distorted AI traffic data (internal traffic)

Exclude internal traffic early to prevent team testing from inflating AI traffic numbers.

1. In GA4, go to Admin → Data collection and modification → Data filters.
2. Create and enable an Internal traffic filter to prevent team testing from polluting your AI traffic checks and Exploration reports.

💡 Pro tip:Set the filter to Testing first, validate it in Realtime or DebugView, then switch it to Active.

3 Distinct Ways to Track AI Traffic

Method 1: Find AI traffic in standard GA4 reports (fastest option)

Start with Traffic acquisition to establish a clear baseline for AI traffic and its core performance metrics.

1. Go to Reports → Acquisition → Traffic acquisition.
2. Use report filters to narrow by Session source/medium, or Session source, to AI referral sources, and search for known AI domains such as chat.openai.com or perplexity.ai.
3. Review quality metrics for these sources, such as Engagement rate, Average engagement time, and Key events/Conversions.

💡 Pro tip:Switch the primary dimension to Session source/medium to capture referral, organic, or less common medium patterns used by some AI platforms.

Method 2: Build an Exploration that filters AI sessions (best for ongoing tracking)

First, set up an Exploration:

1. Go to Explore → Free form or choose Blank.
2. Add dimensions:
2. • Session source/medium (you can also add Session source)
   • Optional: Landing page + query string to understand which pages AI visitors enter from
3. Add metrics: Sessions, Engaged sessions, Key events, and revenue-related metrics if needed.
4. Apply filters:Create a Session segment named “AI referrals” and apply one of the following conditions:
4. • Option A (recommended):Session source matches regex, for example: chat\.openai\.com|perplexity\.ai|gemini\.google\.com|copilot\.microsoft\.com|claude\.ai
   • Option B (if you tag links with UTM parameters you control):Filter by your UTM source, for example: Session source exactly matches chatgpt (or ai_platform)

💡 Pro tip:Filter by Session source (domain), rather than only Session source/medium, because the medium can change, while the domain is usually more stable.

Method 3: Make AI visits easier to attribute (UTM tagging)

For AI-related links you directly manage, UTM tags provide cleaner attribution.

1. When you control the link, such as in newsletters, partner links, or links you share directly, add UTM parameters like these:
1. • utm_source=chatgpt (or perplexity, gemini)
   • utm_medium=ai
   • utm_campaign=ai_visibility
2. In GA4, filter by Session campaign or Session source/medium to cleanly report on tagged AI traffic.

📌 Note:UTM parameters cannot solve links you do not control that appear in AI answers, but for campaigns you do control, they are the most reliable way to avoid being classified as “Direct.”

GA4 can show when AI platforms drive traffic, but it cannot tell you how your brand appears in those AI answers, such as which products are recommended, whether competitors rank ahead of you, or what sentiment the model uses.

WorkDuo closes this gap bytracking brand- and product-level visibility across major AI platformsusing real-time browser tracking, which is more accurate than standard API tracking, so you can identify where changes occur, fix what is losing traction, and then validate the business impact in GA4.

How to Create an AI Traffic Channel Group for Ongoing Reporting

Follow these steps to create an AI traffic channel group for ongoing reporting, so AI visits are clearly categorized and remain visible without manual filtering every time.

Step 1: Go to Channel groups in GA4

You can create and manage custom channel groups at the property level here.

1. In GA4, click Admin。
2. Under Data display, click Channel groups.
3. Click Create new channel group (you need Editor access or higher).

Step 2: Create the channel group

Give the group a clear name so it is easy to select in reports later.

1. Name the group, for example:AI channel group）。
2. If GA4 asks you to confirm the new group, save it.

Step 3: Add a dedicated AI traffic channel

You will create a channel inside the group that captures AI referral traffic through rules.

1. Click Add new channel.
2. Name it (for example: Artificial Intelligence)。
3. Open the channel rule editor.

Step 4: Add matching rules with regex

This is the core step that tells GA4 which sources should be labeled as AI traffic.

1. Set the dimension to Session source。
2. Set the match type to matches regex.
3. Paste a regex list, for example: chat\.openai\.com|perplexity\.ai|gemini\.google\.com|copilot\.microsoft\.com|claude\.ai
4. Click Apply, Save channel, then Save group.

Step 5: Reorder so this channel is evaluated before Referral

GA4 evaluates channels in order, so your AI channel should sit above Referral to capture these sessions first.

1. Click Reorder or the equivalent sorting control.
2. Move Artificial Intelligence above Referral.
3. Click Apply, then Save.

Step 6: Validate the channel group in Traffic acquisition

This confirms that GA4 is now classifying AI traffic under your custom channel group.

1. Go to Reports → Acquisition → Traffic acquisition。
2. Switch the dimension dropdown to yourcustom channel group。
3. Confirm that you can see a row labeled “Artificial Intelligence” with session data.

Step 7: Create an AI-only traffic report and add it to the left navigation

This turns AI traffic into a permanent report your team can access anytime.

1. Open Traffic acquisition and click the pencil icon to customize the report (requires the appropriate permissions).
2. Add a filter so the Session custom channel group exactly matches the Artificial Intelligence channel group.
3. Change the primary dimension to Session source, or Session source/medium, to list individual AI tools.
4. Click Save → Save as a new report, then name it, for example:AI traffic）。
5. Go to Reports → Library, edit the relevant collection, usually Life cycle, drag the new report into the right-side section, then save.

Common Mistakes to Avoid When Tracking AI Traffic in GA4

Before you trust your AI traffic data, watch for these common GA4 setup and reporting mistakes.

1. Ignoring differences in AI referral data

AI traffic data is messy. Some AI tools remove referral source data, use redirects, or pass traffic through changing domains, so some traffic may appear as Direct or Unassigned instead of a clear referral.

💡 Pro tip:Use a regex-based source list in GA4 and update it often.

2. Misreading AI traffic engagement behavior

A common mistake is interpreting AI traffic the same way you would standard search traffic. AI visitors may land on a single page, get the information they need quickly, and leave. If you only look at bounce-like signals, this behavior can appear to underperform.

💡 Pro tip:Review engaged sessions together with landing pages to understand whether the page quickly satisfied user intent or failed to generate engagement.

3. Looking only at traffic volume without checking quality

Growth in AI sessions is useful to see, but volume alone does not tell you whether AI traffic is valuable. You need to review engagement and conversion metrics to understand whether these visits generate leads, revenue, or meaningful actions.

💡 Pro tip:Build a simple AI traffic report that includes sessions, engaged sessions, key events, and conversions in one view, so you always review both volume and quality together.

4. Relying only on GA4 default channel grouping

GA4 default channel grouping is not enough for AI tracking because AI visits are often mixed into Referral, Direct, or Unassigned. A custom channel group is a cleaner, more sustainable way to segment AI traffic.

💡 Pro tip:Create a custom channel and place it above Referral so GA4 classifies AI traffic correctly first.

Measure AI Search Performance End to End with WorkDuo

GA4 can show when AI drives traffic, but it cannot tell youwhyyou appeared, what content was recommended, or when competitors overtook you.

WorkDuo connects AI visibility with GA4 performance by tracking how your brand and products appear across AI platforms, including sentiment and changes over time, then validating real-world impact in analytics. Stop guessing and useStart free trialto see the full picture end to end.

How to Track AI Traffic in GA4: FAQ

1. What is the difference between AI traffic, organic search, and referral traffic in GA4?

In GA4, the difference depends on where the visit came from and how GA4 classifies it, not whether AI was used during the answer process.

Which traffic is considered AI traffic in GA4?

In GA4, “AI traffic” usually refers to visits from AI platforms or AI assistants such as ChatGPT, Gemini, Perplexity, Copilot, and Claude. However, GA4 does not label this traffic as a built-in channel, so these visits often appear under Referral, or sometimes as Direct/Unassigned if referral source data is not passed.

• Example:If chat.openai.com/referral or perplexity.ai/referral appears in Session source/medium, it is usually treated as AI traffic in GA4.

What can and cannot be tracked in GA4?

In GA4, you can track AI traffic only when source/referral data is available. This means you can usually see which AI tool drove the visit through Session source or Session source/medium, such as chat.openai.com/referral, along with sessions, landing pages, engaged sessions, and conversions.

However, you cannot reliably track AI clicks where referral source data has been removed, since they may appear as Direct/Unassigned. GA4 also cannot show AI answer sentiment, share of voice, citation count, or your ranking/visibility in AI results.