How to Track AI Referral Traffic and Improve Marketing Attribution

AI referral traffic is becoming an increasingly important part of the customer journey. People are discovering brands, comparing solutions, and researching products through platforms like ChatGPT, Perplexity, Claude, Gemini, and Microsoft Copilot before ever reaching a website.

But measuring that activity isn't always straightforward.

In this guide, we'll explain how to track AI referral traffic, measure individual platforms like Perplexity, build an AI traffic dashboard, and make more informed attribution decisions.

What Is AI Referral Traffic?

AI referral traffic is website traffic generated when someone clicks a link to your website from an AI assistant or generative AI platform such as ChatGPT, Perplexity, Claude, Gemini, or Microsoft Copilot.

Unlike traditional organic search traffic, these visitors may have already spent significant time researching a topic, comparing products, or evaluating solutions within an AI conversation before clicking through to your website.

For marketers, there are two important categories to understand:

• Identifiable AI referral traffic: Visits where analytics platforms receive enough source or referral information to recognize the AI assistant.
• Unattributed traffic: Visits where sufficient referral information isn't available. These sessions may be classified as Direct or another channel, even if an AI interaction influenced the visit.

That distinction is important. Analytics can tell you what it can observe, but it can't reliably identify the origin of every website visit.

How to Track AI Referral Traffic in GA4

GA4 now includes an AI Assistant channel within its default channel grouping, making it easier to identify traffic from recognized AI assistants. Google added the channel on May 13, 2026, and it requires no setup: when a visit's referrer matches a recognized AI assistant, GA4 sets the medium to ai-assistant and groups the session under AI Assistant.

Two limits are worth knowing up front. Google's definition names sources like ChatGPT, Gemini, Deepseek, Copilot, and Grok, so some platforms, including Perplexity, may still appear under Referral. And clicks from Google's own AI Overviews and AI Mode are not included; GA4 counts them as Organic Search.

Here's how to start measuring it.

1. Check the AI Assistant Channel

Start by reviewing your acquisition reporting in GA4 and looking for traffic attributed to the AI Assistant channel.

This gives you a baseline for the AI referral traffic GA4 can identify.

From there, don't stop at sessions. Look at how that traffic contributes to:

• Engaged sessions
• Landing page performance
• Key events
• Leads
• Purchases
• Revenue

This helps you understand whether AI assistants are simply sending visitors or contributing to meaningful business outcomes.

2. Review Individual AI Referral Sources

Next, break AI traffic down by source to understand which platforms are actually sending visitors.

Depending on the traffic your website receives, you may see referrals associated with platforms such as:

• ChatGPT
• Perplexity
• Claude
• Gemini
• Microsoft Copilot

Looking at individual sources can help you identify which AI platforms matter most to your audience and which website pages are attracting AI-driven visits.

If a platform you expect doesn't show up in the AI Assistant channel, check the Referral channel for its domain before assuming it isn't sending traffic.

3. Analyze AI Traffic by Landing Page

Landing pages provide another important layer of context.

Identify which pages receive measurable AI referral traffic and compare their performance. For example, you may find AI users frequently arrive on:

• Product or service pages
• Comparison pages
• Pricing pages
• Educational guides
• Research or statistics pages
• Detailed how-to content

From there, compare engagement and conversion activity.

This can help your marketing team understand not only where AI traffic is coming from, but what content is successfully moving those visitors forward.

4. Use Custom Channel Groups When You Need More Control

GA4's native AI Assistant classification provides a useful starting point, but some marketing teams may want additional control over how AI traffic is grouped and reported.

Custom channel groups can be used to create reporting definitions around specific sources or classifications. This can be particularly useful when your organization wants to maintain its own AI traffic reporting methodology, separate individual platforms, or include sources the default channel doesn't recognize, such as Perplexity. Google's custom channel group guidance recommends placing an AI channel above Referral so matching sessions are classified correctly.

Whatever approach you use, document your definitions so they remain consistent over time.

How to Track Perplexity Referral Traffic

To check Perplexity referral traffic, don't rely on the AI Assistant channel alone. Perplexity isn't named in Google's current AI Assistant definition, so its visits may be classified as Referral. In the Traffic acquisition report, set the dimension to Session source / medium and search for "perplexity," or build a custom channel group with a rule for perplexity.ai.

Once you've isolated identifiable Perplexity traffic, analyze the same performance indicators you would for any other acquisition source:

• Sessions and users
• Landing pages
• Engagement
• Key events
• Leads or purchases
• Conversion rate
• Revenue, when applicable

For example, your reporting might look like this:

In this example, you can say that GA4 identified 86 sessions associated with Perplexity and nine generated the selected key event.

What you can't conclude is that every unattributed visit to that landing page also came from Perplexity or that you know which individual prompt caused your website to appear in an AI response.

Keeping that distinction clear helps prevent AI reporting from becoming more precise than the underlying data allows.

Why Does AI Referral Traffic Sometimes Show as Direct?

Some visits influenced by AI may ultimately appear as Direct when GA4 doesn't receive enough information to determine their source.

However, Direct traffic is not the same thing as AI traffic.

A Direct session simply means analytics does not have a clearly identifiable referral source. That can happen for many reasons, including:

• Someone typing your URL directly
• Bookmarks
• Missing tracking parameters
• Links from certain documents or applications
• Redirects that remove referral information
• Privacy tools or ad blockers
• Other situations where source information isn't available

That means an increase in Direct traffic can be worth investigating, but it should not automatically be attributed to AI.

Use Direct Traffic as a Diagnostic Signal

If Direct traffic changes significantly, examine:

• Landing pages
• New versus returning users
• Conversion behavior
• Branded search trends
• Campaign activity
• Offline marketing
• Known AI referral patterns

These signals can provide context, but they don't prove that otherwise unattributed traffic came from an AI assistant.

Can AI Referral Traffic Be Trusted for Marketing Attribution?

AI referral data can be trusted for measuring traffic your analytics platform can actually identify, but it should not be treated as a complete measurement of AI's influence on the customer journey.

A useful way to approach the data is to separate what you know from what you infer.

This distinction becomes especially important when reporting AI performance to clients, executives, or other stakeholders.

AI referral reporting should show the measurable impact of the channel without presenting assumptions as confirmed attribution.

Can Server Logs Help Track AI Traffic?

Server-side and CDN logs can provide another useful source of information, particularly when analyzing AI crawler activity and requests that aren't captured by JavaScript-based analytics.

They can help marketing and technical teams better understand how AI platforms interact with website content.

However, server logs don't automatically reveal the original acquisition source of a human website session when that referral information was never transmitted.

For that reason, server logs should be treated as supplementary evidence, not a way to definitively reclassify every unattributed visit as AI traffic.

What Should an AI Referral Traffic Dashboard Include?

An AI referral traffic dashboard should show identifiable AI traffic by source, landing page, engagement, conversions, and reporting period.

At minimum, consider including:

The goal isn't simply to prove that AI traffic exists. Your dashboard should help answer more useful questions:

• Which AI platforms are sending visitors?
• What content are those visitors discovering?
• What do they do after arriving?
• Is that activity contributing to business results?

How to Monitor AI Referral Traffic Over Time

AI referral traffic is evolving quickly, so a single snapshot won't tell you much about its long-term impact.

Create a repeatable measurement process:

1. Use a consistent AI traffic definition. Decide what qualifies as identifiable AI traffic and document it.
2. Track traffic by AI platform. Monitor ChatGPT, Perplexity, Gemini, Claude, Copilot, and other relevant sources separately when possible.
3. Compare consistent reporting periods. Use equal timeframes for month-over-month or year-over-year analysis.
4. Monitor landing pages. Identify which content consistently earns AI referrals.
5. Measure conversions. Don't evaluate AI performance on sessions alone.
6. Document tracking changes. Record changes to GA4 classifications, custom channel groups, or reporting methodology. Note the date the AI Assistant channel began populating in your property, since the channel isn't retroactive and earlier AI visits remain under Referral.
7. Compare AI with other channels. Look at engagement, conversion rate, revenue, and other business outcomes alongside traffic volume.

Consistency matters. If you change how AI traffic is defined from one reporting period to the next, an apparent increase or decrease may reflect a tracking change rather than a real shift in performance.

Common AI Traffic Attribution Mistakes

As AI traffic becomes a standard part of marketing reporting, avoid these common attribution mistakes:

Why AI Referral Traffic Matters

AI referral traffic may still represent a relatively small percentage of overall website traffic for many organizations, but traffic volume isn't the only measure of a channel's value.

Users who arrive through AI assistants may already have researched their problem, evaluated possible solutions, or compared brands before visiting your site.

The data supports that. According to Adobe Analytics, traffic from AI sources to U.S. retail sites grew 393% year over year in the first quarter of 2026, and in March 2026, AI-referred visits converted 42% better than non-AI traffic from channels such as paid search and email.

That's why marketers should evaluate AI traffic based on the outcomes it produces, not simply how many sessions it generates.

As AI-assisted discovery grows, maintaining a reliable baseline now also gives your team historical data to compare against later.