How to Track AI Referral Traffic in GA4: A Guide for Marketing Agencies

Your client's website traffic looks normal. Organic traffic and enquiries from paid campaigns as well as direct visitors are continuing to remain steady. At a monthly review it is noticed that referrals from ChatGPT and Perplexity have been made.

To which parts of the marketing report will these visitors go?

Another type of traffic to watch that agencies need to know about is people who are researching products, comparing services or finding websites.

Agencies that track AI referral traffic in GA4 can help marketing teams determine whether there are visits from AI platforms and analyse how these visitors are interacting with a website.

But not all AI-driven visits are detectable. Some sessions may be classified as different traffic, and care needs to be taken in interpreting them.

Agencies should identify, measure and report relevant referrals.

What Is AI Referral Traffic?

AI referral traffic is the traffic that arrives from sites that are connected with AI channels.

Users can ask ChatGPT to recommend project management software, for example, What project management software would you suggest for this project? Google Analytics 4 (GA4) could record that visit if it includes a link to a website that the person clicks on in the response.

AI platforms worth investigating include:

• ChatGPT: Visitors following links shared in AI conversations.
• Perplexity: Users opening websites referenced in AI-powered answers.
• Claude: Visitors following links provided during conversations.
• Microsoft Copilot: Users visiting websites through AI-assisted experiences.
• Gemini: Users accessing links via Google's AI tool.

Not all of these platforms pass the referral information in the same manner.

Most importantly, GA4 tracks website activity and not every time a brand is mentioned, recommended or interacted with in an AI platform.

Where Does AI Traffic Appear in GA4's AI Assistant Channel?

GA4 uses acquisition dimensions to categorise visits to its sites.

Since May 2026, GA4 has included an AI Assistant default channel for recognised AI traffic sources. Agencies should check this channel alongside individual session sources rather than relying only on Referral traffic.

AI referrals may appear in the AI Assistant or Referral channel but the classification can vary based on what information is received and the rules used by GA4's attribution system.

For instance, a visit that is linked to ChatGPT could be of the form chatgpt.com / referral.

However, this is not guaranteed.

Do not assume that all the traffic from the AI-related session comes from a specific channel; please verify with the 'session sources' of the recorded session.

A Practical Workflow for Identifying AI Referrals

Agencies can explore identifiable AI traffic in the existing GA4 acquisition reports and exploration features.

Step 1: Open the Traffic Acquisition Report

Log in to the respective GA4 property and navigate to Traffic acquisition report in Reports.

Choose a reporting period that is appropriate and is a month that has just ended.

This will give you an overview of website sessions by acquisition source.

Step 2: Examine Individual Session Sources

If available, change the primary dimensions of the report to Session source / medium or Session source.

Look for recognisable AI-related domains such as:

• chatgpt.com
• perplexity.ai
• claude.ai
• copilot.microsoft.com
• gemini.google.com

These are just examples of where you can find these; the list isn't exhaustive. Recorded domains will change depending on platforms and referral behaviour.

Step 3: Create a Focused Exploration

Use GA4's Explore section to get more in-depth analysis.

Develop an exploration to include session-based acquisition dimensions and the appropriate engagement metrics.

Use filters to isolate the recognisable AI sources within the property.

This means that agencies don't have to sift through irrelevant traffic again and again.

Step 4: Review Landing Pages

Determine which web pages are getting traffic from AI websites.

For instance, an agency may find that articles comparing their service with others deliver more identifiable AI new business to their website than their homepage.

This may indicate that some content types should be further explored.

But performance of the landing page is not enough to prove that the reason why an AI assistant referred to a specific website.

Step 5: Analyse Engagement and Key Events

But it's not all about traffic volume.

Analyse if the visitors are interacting with content, navigating the site and completing configured key events.

A smaller referral source that provides good quality enquiries may be worth more than a larger source that provides few or no meaningful interactions.

Which Metrics Should Agencies Include in Client Reports?

When agencies track AI referral traffic in GA4, they should combine acquisition data with engagement and business outcomes.

Agencies should compare the same period of time and describe limitations of small samples.

For instance, an increase of five sessions to ten sessions is a significant percentage increase, but that is still not a large number of sessions to make large decisions.

Why AI Referral Numbers May Be Incomplete

One of the biggest reporting challenges is that identifiable referrals do not represent the complete influence of AI platforms.

There are multiple factors that impact the information that is available in GA4.

Missing referral details: If there are visits that have been sent without adequate information about the original source.

Direct traffic classification: Sessions with no data available for acquisition can be categorised as Direct, and the GA4 attribution system may also allocate credit to previous known sources.

Browser and privacy restrictions: Consent preferences and browser behaviour / privacy restrictions may impact measurement.

Different AI experiences: Links, redirects and referrals can be managed in different ways by platforms.

AI features within Google Search: AI-powered visits from Google Search experiences might be counted as part of organic search traffic instead of as referrals from AI traffic.

Therefore, agencies must not take it for granted that all Direct sessions were created by an AI platform.

Nor can they report on the number of identifiable AI referrals as a percentage of the number of visitors impacted by AI recommendations.

Turn Referral Data Into Useful Client Insights

There are three practical questions which should be addressed in a good agency report.

Which Platforms Are Sending Visitors?

Find and review the AI-related sources in GA4.

Don't claim minor differences in platforms as significant rankings until there's enough data.

Which Pages Attract Those Visitors?

Review landing pages to identify traffic sources.

Identify patterns in educational articles, service pages, product comparisons and other kinds of content.

Are Visitors Taking Valuable Actions?

Compare with other acquisition sources, engagement and configured key events.

Where relevant, include qualified enquiries, and sales from the client's current records.

This enables clients to determine if measurable AI referrals work in favour of their goals or just drive up web traffic.

Common Mistakes Agencies Should Avoid

Even helpful analytics can be misinterpreted if there are significant differences that are not captured.

Be careful not to make these common reporting errors:

• Treating all Direct traffic as AI-generated visits.
• Assuming every AI citation produces a website click.
• Presenting referral traffic as complete AI visibility data.
• Comparing different acquisition channels without consistent definitions.
• Claiming that referral growth proves stronger AI recommendations.
• Making strategic decisions from very small traffic samples.

AI visibility is a concept that is linked with identifiable referrals and business outcomes; however, they are not synonymous.

If they are separated, their reports are more credible.

Make AI Referral Monitoring Part of Regular Reporting

While AI discovery can be another helpful space for agencies to explore, it shouldn't be used as a substitute for traditional SEO and analytics reporting.

Start by describing measurable referrals, breaking down their traffic and aligning the available data with relevant business activities.

Communicate what the data supports and what is not clear, most importantly.

At DM Cockpit, we know agencies can face some hurdles during their marketing performance review across different channels, and we are here to help them overcome these challenges! Our digital marketing dashboard brings together analytics and reporting capabilities to aid in much better, more organised client reviews and informed decisions.

Frequently Asked Questions

1. Can GA4 automatically identify all traffic from AI platforms?

No. GA4 can identify some visits when suitable source information is available. However, missing referral details, attribution rules and tracking limitations mean that not every AI-influenced visit can be recognised.

2. Is ChatGPT traffic considered organic or referral traffic?

Identifiable visits from ChatGPT may appear as referral traffic in GA4. However, the recorded source and channel depend on the information available and GA4's attribution rules.

3. Can agencies create a separate AI traffic report in GA4?

Yes. Agencies can use GA4 explorations, filters and available acquisition dimensions to examine known AI-related sources. Such reports should be labelled as identifiable AI referrals rather than complete AI-generated traffic.

4. Which GA4 metrics are most useful for analysing AI visitors?

Sessions, engaged sessions, engagement rate, average engagement time, landing pages and key events provide a useful starting point. Together, they help agencies evaluate traffic volume and visitor behaviour.

5. Why might AI referrals appear as Direct traffic?

Some visits may lack usable referral or campaign information because of browser behaviour, redirects or other tracking limitations. However, agencies cannot assume that all Direct traffic originates from AI platforms.

6. Does increasing AI referral traffic mean better AI search visibility?

Not necessarily. Higher referral numbers indicate more recorded visits from identifiable sources. They do not independently prove stronger AI visibility, increased brand recommendations or improved business performance.