How to Track AI Traffic in GA4

 AI tools like ChatGPT, Perplexity, Claude and Gemini are changing how users find and engage with websites. If you’ve noticed unusual spikes in direct traffic or certain landing pages generating conversions without a clear source, there’s a strong possibility that traffic is coming from AI platforms.

GA4 doesn’t highlight these sessions in standard reports and typically attributes them to referral traffic. Without a custom setup, it’s easy to miss whether your referral traffic sources are from AI search engines, or other sites linking to yours. In a nutshell, it’s not ideal.

Fortunately, there’s a simple way to set up a custom report in Google Analytics 4 to identify AI website traffic using a regex filter that captures visits from tools like ChatGPT, Gemini, Microsoft Copilot and any other AI search tools you want to track. Once you’ve got this set up, you’ll have clear visibility into which AI platforms are sending users to your site, and how those users are interacting with your content.

Let’s get into it!

Common AI Traffic Sources in 2025

AI tools that link to your site are usually classed as referral traffic in GA4’s Traffic Acquisition report. This means they’re mixed in with standard referrers in your default reports, making it harder to separate AI-driven visits from traditional link sources.

Below is a list of the most common AI platforms that send traffic to websites in 2025 — and that you’ll want to include in your custom tracking setup:

• ChatGPT – chat.openai.com
  OpenAI’s flagship tool sends traffic when users click links directly from a chat session.
• OpenAI – openai.com
  Occasionally appears separately from ChatGPT, especially when content is accessed via shared outputs or embedded views.
• Google Gemini – gemini.google.com
  Google’s AI assistant that passes referral data when users visit recommended pages.
• Perplexity – www.perplexity.ai
  A search-style AI that regularly cites and links out to web pages. One of the most trackable AI traffic sources.
• Claude – claude.ai
  Anthropic’s AI assistant and used across various platforms. Sends traffic when links are shared in responses.
• Microsoft Copilot – copilot.microsoft.com
  Built into Microsoft 365 and Bing, and often passes referrer data when surfacing website links.
• Edge Services – edgeservices.bing.com
  Typically appears when links are clicked via Bing AI or Copilot in the Microsoft Edge browser.

You may want to track additional AI tools depending on your industry and audience. In the next section, we’ll show you how to set up a GA4 report that isolates traffic from these platforms and how to add in others as needed.

How to Track AI Traffic in GA4

To track traffic from AI chatbots properly, you need to set up a custom report in GA4 that filters traffic by session source. This allows you to isolate visits from platforms like ChatGPT, Gemini and Perplexity, so you can monitor how AI referral traffic behaves on your site.

Follow the steps below to set it up:

1. Create a New Custom Report

• In your GA4 property, go to Library at the bottom of the left-hand navigation
• Click Create new report and choose Create detail Report
• Start with a Blank report to build from scratch

2. Add Dimensions

Add the following dimensions to the report so you can see where the user came from and which page they landed on:

• Session source
• Landing page

3. Add Metrics

Include the key metrics you want to analyse:

• Sessions
• New users
• Engaged sessions
• Engagement rate
• Key events

You can also include additional metrics depending on how you measure conversions or performance.

4. Apply a Filter for AI Traffic Sources

Now, isolate the report to AI-driven traffic by setting up a filter with these settings:

• Dimension: Session source
• Match type: Matches regex
• Value:

.*chatgpt.com.*|.*perplexity.*|.*edgeservices.*|.*copilot.microsoft.com.*|.*openai.com.*|.*gemini.google.com.*|.claude.ai.*

This will capture visits from all the common AI platforms covered earlier. You can expand this filter by adding any new domains you notice appearing in your referral traffic.

5. Save Your Report

Once you’re happy with the metrics and dimensions you can see, give your report a clear name, like “AI Traffic Overview”, and save it to your report library.

What to Look for in Your AI Search Traffic Data

Once your AI traffic report is live, the next step is to interpret what the data is actually telling you. Here’s what to focus on.

1. Traffic Volume by AI Tool

Check which AI tools are your biggest sources of traffic. ChatGPT and Perplexity tend to be the most active, but this will vary depending on your industry and content type. This helps you prioritise which platforms to pay attention to.

2. Top Landing Pages

Switch your dimension to Landing Page to see which pages are getting traffic from AI tools. This helps you identify content being recommended, cited or linked in AI responses and spot new content opportunities based on what’s already working.

3. User Engagement

Metrics like Engaged Sessions, Engagement Rate and Key Events help you analyse how AI-referred users are interacting with your site compared to other sources.

For example, if AI traffic shows strong engagement on certain pages but low conversions, it may point to an intent mismatch or an opportunity to optimise for that audience.

Alternatively, a specific landing page may have much higher conversions from AI tools compared to Google search, indicating AI tools are able to match your content with your target audience better than traditional search engines.

4. New vs Returning Users

Add the Returning Users metric to compare with the New Users metric to understand whether AI traffic is driving net-new audiences or helping users return through new discovery paths.

5. Spikes or Patterns

Keep an eye on any sudden spikes in AI traffic. This can happen if:

• A piece of content starts trending in AI answers
• An AI tool updates its browsing or citation behaviour
• Your domain is being used as a reference in popular prompts

If you see a spike, dig into which platform caused it, what pages they are sending traffic to and then review your sales funnel to make sure you’re making the most of the traffic spike.