How to See AI Referral Traffic in Google Analytics 4

Chances are, a new category of referral traffic is showing up in your analytics, and you might be missing it. AI search assistants like Perplexity and ChatGPT are increasingly sending high-intent visitors to websites by citing sources in their answers. This article will show you exactly how to find and analyze your AI referral traffic in Google Analytics 4, so you can measure its impact and optimize your content for this emerging channel.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

Why Tracking AI Referral Traffic Is A Big Deal

For years, marketers have obsessed over referral traffic from social media, forums, and other blogs. Now, AI is a new player on the scene, and it acts differently. Unlike a casual click from a social media post, traffic from an AI assistant often comes from a user who has asked a very specific question. When the AI uses your article as a numbered source to answer that question, the resulting click-through is from someone actively seeking the expertise you offer. This is high-quality, high-intent traffic.

Understanding this traffic helps you answer critical questions:

• Which AI platforms are valuable sources of traffic for my business?
• What specific content on my site is being used as a source by AI?
• Are these visitors engaging with my content and converting?
• How can I create more content that AI models will want to cite?

By monitoring this traffic, you can start adapting your content strategy to leverage this new discovery channel instead of just watching it pass by.

How to Find AI Referral Traffic in GA4 (The Quick Way)

If you just want a quick snapshot of what AI sources are sending you traffic, you can use the standard Traffic Acquisition report in Google Analytics 4. It only takes a minute.

Step 1: Go to the Traffic Acquisition report

In the left-hand navigation of GA4, go to Reports > Acquisition > Traffic acquisition.

Step 2: Change the Primary Dimension to "Session source"

By default, this report is grouped by the "Session default channel group." To see the specific websites sending you traffic, click the drop-down arrow on the primary dimension and select Session source.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

Step 3: Filter for Referral Traffic

Now, we want to look only at referral traffic. Use the search box right above the data table and type "Referral." This will filter the results to show you only the sessions that GA4 has categorized into the referral channel group. Your table should now list out all the different websites sending you traffic.

Step 4: Identify AI Sources

Scan through the list of session sources. You're looking for URLs associated with popular AI tools. The most common ones you'll see are:

• perplexity.ai
• chat.openai.com
• gemini.google.com (Google Gemini, formerly Bard)
• copilot.microsoft.com
• claude.ai
• bing.com (If you previously separated out Bing organic traffic, a lot of traffic from Copilot can still be assigned to here. It can be tricky, but look for anomalous spikes in Bing referrals)

If you have a lot of referral sources, you can get more specific with the search bar. Instead of "Referral," try searching for "perplexity" or "openai" to instantly see if they appear in your reports.

Creating a Custom "AI Referral" Report in GA4

The method above is useful for a quick check, but it’s not ideal for ongoing analysis. You have to re-apply the filters every time, and you can’t easily compare AI sources against each other. For that, you need a custom report in the "Explore" section of GA4.

This might seem advanced, but it's straightforward if you follow these steps. Once you build it, you'll have a saved, reusable report you can check anytime.

Step 1: Create a New "Free form" Exploration

Head to the Explore section in the left navigation and click on Blank exploration or select the Free form gallery option.

Step 2: Add Your Dimensions and Metrics

In the "Variables" column on the left, you need to import the dimensions (the "what") and metrics (the "how many") you want to analyze.

Click the + icon next to Dimensions and add the following:

• Session source
• Landing page + query string
• Country
• Device category

Click "Import." Now do the same for Metrics. Click the + and add:

• Sessions
• Total users
• Engaged sessions
• Average engagement time
• Conversions

Click "Import" again. Your "Variables" panel should now be populated with these options.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

Step 3: Build the Report Table

Now, you need to drag and drop your imported variables into the "Tab Settings" column.

• Drag Session source from "Dimensions" into the Rows area.
• Drag your desired metrics (like Sessions, Engaged sessions, Average engagement time, and Conversions) from "Metrics" into the Values area.

You'll immediately see a report appear on the right, showing these metrics for all your session sources.

Step 4: Filter for AI Sources (The Magic Step)

This is where we tell GA4 to only show us the data we care about. In the "Tab Settings" column, find the Filters box.

1. Drag Session source here or click and select it from the list.
2. Set the condition to matches regex. This sounds technical, but it’s just a powerful way to list multiple conditions at once. Regex allows you to use a vertical bar | to mean "OR".
3. In the text box, enter the names of the AI sources you want to track, separated by |. For example:

perplexity.ai|chat.openai.com|gemini.google.com|copilot.microsoft.com

1. Click Apply.

Your report will now be filtered to show traffic coming only from those AI platforms you specified. Give your exploration a name at the top left, like "AI Referrals Dashboard," so you can easily find it later.

How to Analyze and Get Insights from Your AI Referral Traffic

Having the data is one thing, using it to make better decisions is another. With your new custom report in place, you can start digging for insights.

Which AI Platforms Are Driving the Most Value?

Your report immediately shows you which AI sources are sending the most Sessions and Users. But don't stop there. Look at the engagement metrics.

You might find Perplexity.ai sends less traffic than Gemini, but its Average engagement time is double, indicating that visitors are more absorbed in the content. Add your key conversion events to the Conversions column to see which sources are actually driving valuable outcomes, like lead form submissions or sales.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

What Content Is Getting Surfaced By AI?

This is arguably the most valuable insight you can get. In your custom report, add Landing page + query string as your Rows dimension, and keep Session source as a Columns dimension. This will create a pivot table showing exactly which URLs on your site are receiving traffic from each AI platform.

This reveals the content that AI trainers and models deem authoritative and useful. It's a roadmap for what to create next. If you see your article "10 Ways to Fix a Leaky Faucet" is getting tons of traffic from AI referrals, it's a strong signal to create more in-depth content around home repair.

Are These Visitors Sticking Around?

Look at the Engaged sessions metric. This tells you the number of sessions that lasted longer than 10 seconds, had a conversion event, or had at least 2 pageviews. If this rate is low for a particular AI source, it might mean the clicks are accidental or the content doesn't align perfectly with the user's initial query.

Future-Proofing Your Content for an AI-Driven World

Identifying this traffic is the first step. The next is optimizing for more of it.

• Answer Questions Clearly: Structure your content logically. Use clear headings (H2s) in the form of questions (How do you fix...?, `What is the best...?) and then answer them directly below. Many AI answers are extractive, meaning they pull sentences or paragraphs directly from a source.
• Focus on E-E-A-T: Double down on Expertise, Experience, Authoritativeness, and Trustworthiness. Publish original data, feature case studies, and cite credible external sources. These are the signals that both traditional search engines and AI models are looking for.
• Use Structured Data: Implementing schema markup like FAQPage or HowTo helps machines understand your content's structure and purpose more easily, making it a more efficient sourcing option for them.

Final Thoughts

Tracking referral traffic from platforms like Perplexity, ChatGPT, and Gemini is a non-negotiable part of modern marketing analytics. By checking your standard GA4 reports or building a simple custom exploration, you can finally put hard numbers to this growing channel and uncover which AI tools - and which pieces of your content - are most valuable to your business.

As you move beyond manual report building, the next step is talking to your data directly. Instead of building these reports and fiddling with filters in GA4, we prefer connecting our data sources and just asking for insights. Using Graphed, we can connect Google Analytics and instantly ask in plain English: "Show me a bar chart of sessions from all AI referral sources this quarter and what landing pages they visited." It builds the charts instantly, letting us spend our time analyzing the data instead of just chasing it down.