How to Track AI Traffic in GA4 in 2026

Track AI traffic in GA4 correctly in 2026 and you can see more AI-referred visits without relying entirely on custom regex filters or manually built channel groups.

Google Analytics now includes an AI Assistant channel for recognised referrals from AI assistants such as ChatGPT, Gemini and Claude. That gives marketers a much cleaner starting point for measuring AI-driven website traffic.

However, GA4 still does not measure every interaction a potential customer has with AI search. Someone may discover your brand in ChatGPT, Gemini, Perplexity, Google AI Overviews or Google AI Mode without clicking through to your website at all.

That means AI referral traffic and AI visibility need to be measured separately.

This guide explains how to track AI traffic in GA4 using the native AI Assistant channel, Session source / medium, Explorations, custom channel groups and UTMs where appropriate. It also shows how to connect AI traffic with key events, CRM opportunities and pipeline rather than treating sessions as the final KPI.

If your broader goal is to improve visibility inside AI-generated answers, start with our guide to Generative Engine Optimisation and our broader AI Search SEO guide.

What Is AI Traffic in GA4?

AI traffic in GA4 is website traffic that Google Analytics can associate with an AI-driven source or assistant.

When GA4 recognises an AI-assistant referral, it can group the session under the AI Assistant channel. You can then use dimensions such as Session source, Session source / medium, landing page and page referrer to understand which platforms are sending visitors and what those visitors do after reaching your website.

This gives marketers a much cleaner way to separate identifiable AI referrals from traditional Referral, Organic Search and Direct traffic.

But there is an important limitation.

GA4 records website activity. It does not tell you every time your company appears, gets cited or is recommended inside an AI answer.

That creates two different measurement layers.

AI referral traffic measures identifiable visits that reach your website from AI platforms.

AI visibility measures mentions, citations, source links and brand presence inside AI-generated answers, including interactions that produce no website click.

For the second layer, use a dedicated AI measurement framework such as our guide to AI Search Visibility Metrics.

Which AI Platforms Can You Track in GA4?

The exact source information available depends on how the AI platform sends the visit and whether usable referral information reaches your website.

AI platform	Source or referrer to inspect	Recommended treatment
ChatGPT	chatgpt.com and historical OpenAI referral values	Check AI Assistant first, then Session source / medium
Gemini	gemini.google.com	Analyse as AI-assistant referral when recognised
Perplexity	perplexity.ai	Check AI Assistant and raw source data
Claude	claude.ai	Analyse as AI-assistant referral when recognised
Microsoft Copilot	Microsoft or Copilot-related source values may vary	Inspect source before manually classifying
Other AI assistants	Varies by platform	Review Session source / medium before adding custom rules
Google AI Overviews	Google Search	Do not automatically classify as third-party AI referral traffic
Google AI Mode	Google Search	Analyse separately from external AI assistants

Do not treat this as a permanent master list. AI products change domains, redirect behaviour, applications and referrer handling.

Your actual GA4 data should remain the source of truth.

If you want to compare how the major platforms behave from a B2B SaaS perspective, see our analysis of ChatGPT vs Perplexity vs Gemini vs Claude.

How to Track AI Traffic in GA4 in 2026

The most important change in 2026 is that you should start with GA4’s native reporting before building custom tracking.

The basic workflow is:

Reports → Acquisition → Traffic acquisition

Look at Session default channel group and check whether AI Assistant appears.

Then use Session source / medium to understand which individual sources are contributing traffic.

This native reporting layer should now be the starting point for most websites.

Custom regex filters, Explorations and channel groups are still useful, but they should solve specific reporting gaps rather than act as the default setup.

Step 1: Check the Native AI Assistant Channel

Open your GA4 property and navigate to:

Reports → Acquisition → Traffic acquisition

Review the Session default channel group dimension and look for AI Assistant.

Then add Session source / medium as a secondary dimension where useful.

This lets you answer questions such as:

Which AI platforms send the most visitors?

Which landing pages receive AI referral traffic?

Do AI visitors engage differently from organic visitors?

Which AI sources generate key events?

Does AI traffic contribute to qualified leads or pipeline?

For B2B SaaS, that final question is usually more important than session volume.

Break AI Traffic Down by Platform

The AI Assistant channel gives you the overall category, but Session source / medium provides the source-level detail needed for useful analysis.

Compare platforms using metrics such as sessions, engaged sessions, engagement rate, landing pages, key events and session key event rate.

This matters because AI traffic is not one homogeneous acquisition channel.

A visitor arriving from ChatGPT after asking for a vendor comparison may have very different intent from someone clicking through from Gemini after researching a broad educational topic.

Do not assume the platform sending the most sessions is automatically the platform producing the most pipeline.

Default Channel Group vs Custom Channel Group

You no longer need to create a custom “AI Traffic” channel group simply to isolate every recognised AI-assistant referral.

GA4’s Default Channel Group should be the first reporting layer.

A custom channel group can still be useful when you need greater control over how AI traffic is classified.

For example, you may want to separate:

ChatGPT traffic
Gemini traffic
Perplexity traffic
Claude traffic
other AI-assistant traffic

You may also need a custom group when a source appears in GA4 but is not classified in the way your organisation expects.

Custom channel groups can be managed under:

Admin → Data display → Channel groups

Be careful with rule order. If several conditions could match the same traffic source, classification logic can produce unexpected results.

For most teams, the cleaner workflow is:

Native AI Assistant channel → source-level analysis → Exploration → custom channel group only when necessary.

Use GA4 Explorations for Deeper AI Traffic Analysis

GA4 Explorations are useful when the standard Traffic acquisition report does not provide enough detail.

Go to:

Explore → Free form

Useful dimensions include Session source / medium, Session default channel group, Landing page + query string and Page referrer.

Useful metrics include Sessions, Engaged sessions, Engagement rate, Key events and Session key event rate.

Start by filtering for the native AI Assistant channel.

You can then create a second view using Session source / medium or Page referrer to investigate traffic that GA4 may not be classifying exactly as you expect.

A working regex might include sources such as:

chatgpt\.com|chat\.openai\.com|perplexity\.ai|gemini\.google\.com|claude\.ai

Treat this as a maintainable reporting rule rather than a permanent universal definition. Review the source values actually appearing in your property before changing the regex.

If you want to turn this analysis into a reusable reporting view, follow our step-by-step guide to creating a custom report in GA4.

Do You Need Custom GA4 Events for AI Traffic?

Usually, no.

You do not need a special ai_chatbot_click event just to identify people arriving from AI assistants.

Acquisition dimensions should handle the traffic-source information.

Instead, your events should measure what the visitor does after arriving on the website.

Examples might include a lead form submission, signup, demo request, trial activation, pricing-page interaction or another meaningful action associated with your funnel.

The most important events can then be treated as key events.

This produces a much more useful question:

How many AI-referred sessions produced important business actions?

That is far more valuable than simply counting AI-related event fires.

Should You Use UTM Parameters to Track AI Traffic?

UTM parameters can help with AI-related traffic when you control the exact URL being distributed.

For example, UTMs may be useful in a controlled AI-powered product experience, partner distribution arrangement, campaign or another environment where you determine the destination link.

A structured URL might use parameters such as:

utm_source
utm_medium
utm_campaign

But UTMs should not be treated as a universal method for measuring public AI citations.

You generally do not control how ChatGPT, Gemini, Perplexity or another public AI platform cites your website.

The assistant may link to your canonical page, another URL, or a URL without your campaign parameters.

Therefore:

UTMs complement AI referral tracking. They do not replace source and referrer measurement.

AI Referral Traffic Is Not the Same as AI Visibility

GA4 answers an important but narrow question:

Which identifiable AI-driven visits reached the website?

It cannot fully answer:

How visible is the brand inside AI-generated answers?

That distinction matters because a large part of AI discovery can happen without a click.

A potential buyer might ask an AI assistant about software in your category, see your company recommended, continue researching and eventually reach your website through branded Google Search or Direct traffic.

GA4 cannot reliably reconstruct that entire journey from the final website visit.

Your reporting should therefore separate:

Referral measurement: AI sessions, sources, landing pages, engagement and key events.

Visibility measurement: AI mentions, citations, cited URLs, prompt coverage and share of voice.

For a full measurement framework, use our AI Search Visibility Metrics guide.

Google’s own AI search experiences also require different thinking from third-party chatbot referrals. Our guide to Google AI Mode explains how that Search experience differs from tools such as Gemini and ChatGPT.

If Gemini is strategically important to your market, you can separately monitor mentions and citations using the framework in our Gemini Visibility Tracker.

AI Traffic Attribution Limitations

When you track AI traffic in GA4, treat the reported number as observable referral traffic rather than the total influence of AI on your buyers.

Referral information can be lost or unavailable for several reasons.

A user might copy a URL from an AI conversation and paste it into a browser.

An app or browser may not pass complete referral information.

A redirect may interfere with attribution.

The user may first discover your brand through an AI response but return days later through Google or Direct traffic.

In these situations, the original AI interaction may not appear as an AI Assistant session.

There is also a difference between understanding how a session was acquired and understanding which touchpoints contributed to a later conversion.

For acquisition analysis, session-level dimensions such as Session source / medium are generally the most useful starting point.

For commercial analysis, compare that data with key events, CRM source data and your wider attribution model.

Troubleshooting AI Traffic Tracking in GA4

Problem	Likely explanation	What to check
AI Assistant does not appear	No recognised AI referral traffic in the selected period	Expand the date range and inspect Session source / medium
ChatGPT or another AI source appears as Referral	GA4 classification differs from your reporting expectation	Inspect raw source data before creating a new rule
AI-influenced traffic appears as Direct	Referral information may not have reached GA4	Treat the AI Assistant total as measurable traffic, not total influence
Custom AI channel is too large	Regex or channel rules may be too broad	Review every included source
Custom AI channel is too small	Relevant sources may be missing	Inspect actual source values before expanding the regex
Sessions appear but no conversions do	Visitor intent or conversion tracking may be weak	Review landing pages and key events
CRM shows an AI lead but GA4 does not	The buyer may have returned in another session or attribution information was lost	Compare first-touch, recent-touch and self-reported source data

Behavioural metrics can help diagnose the quality of AI-referred visits, but they should not be used as proof that traffic is automated or invalid.

If you need more context on engagement analysis, see our guide to finding and interpreting bounce rate in GA4.

Measure AI Traffic Against Key Events

Once you can track AI traffic in GA4, determine whether the traffic contributes to meaningful actions.

Depending on your business, the important outcomes might include demo requests, contact submissions, trial registrations, account creation or another high-intent conversion.

Then compare AI Assistant traffic with Organic Search, Referral, Paid Search and your other major channels.

The metrics worth watching include:

AI Assistant sessions
engaged sessions
engagement rate
key events
session key event rate
landing pages
qualified leads
revenue where applicable

For B2B SaaS, however, even a lead is not the final measurement layer.

Connect AI Traffic to CRM and Pipeline

The most useful AI traffic report for a B2B SaaS company should eventually extend beyond GA4.

Where your consent and analytics architecture permit it, preserve acquisition information in your CRM so leads can be followed through the revenue funnel.

Useful fields might include original source, recent source, landing page, campaign information and the conversion associated with the lead.

Then evaluate AI-driven acquisition against qualified leads, sales opportunities, pipeline and closed revenue.

For example, one AI platform might send hundreds of low-intent visits while another sends relatively few sessions but generates a meaningful number of qualified opportunities.

In that scenario, session volume alone would produce the wrong conclusion.

This is why AI measurement should fit into a wider data-driven marketing strategy for B2B SaaS.

The same source-to-revenue principles used in paid acquisition can also help structure this measurement. Our guide to connecting marketing spend and acquisition data to pipeline covers the broader attribution architecture.

Simple AI Traffic Reporting Template

A practical reporting view can combine GA4, AI visibility and CRM metrics without forcing everything into one artificial score.

KPI	What it tells you
AI Assistant sessions	Identifiable AI referral volume
AI sessions as a percentage of website sessions	Relative acquisition contribution
Sessions by AI source	Platform-level traffic
AI landing pages	Which content earns AI-driven clicks
Engagement rate	Initial traffic quality
Key events	Important onsite actions
Session key event rate	Conversion efficiency
Leads	Lead-generation contribution
Qualified leads	Lead quality
Opportunities	Sales contribution
Pipeline	Commercial impact
Revenue	Closed business where measurable
AI mentions and citations	Visibility that may not result in a click

For executive reporting, one useful funnel is:

AI Assistant sessions → key events → qualified leads → opportunities → pipeline

Keep AI citations and mentions beside this funnel as a separate visibility layer.

Do not add citation counts directly to sessions or create a combined “AI score” unless the methodology has a clear business reason.

How to Analyse AI Traffic Performance

Start with the landing pages receiving AI referrals.

Identify which content formats, topics and buyer stages are attracting visits from AI platforms.

Then compare those pages with your organic search winners.

A page that performs strongly across Organic Search and AI referrals may be demonstrating useful topical authority or strong buyer-intent alignment.

You should still continue to check website keyword rankings alongside AI referrals because traditional search visibility remains an important part of the discovery journey.

Next, compare AI platforms individually.

Do not automatically conclude that the platform sending the most traffic deserves the most attention.

Ask whether those visitors engage, convert and enter the pipeline.

Common Mistakes When Tracking AI Traffic in GA4

One of the most common mistakes in 2026 is starting with a complicated custom regex before checking GA4’s native AI Assistant reporting.

Another is assuming AI Assistant traffic represents total AI visibility. It does not. It measures identifiable referral sessions, not every mention or citation.

Avoid broad custom regex rules that accidentally pull ordinary referral traffic into your AI category.

Do not create unnecessary custom GA4 events to duplicate information that already exists in acquisition dimensions.

Do not judge AI platforms solely by visits. Compare their traffic with meaningful key events and commercial outcomes.

Finally, if you already have a historical custom AI channel, document how it was defined before replacing it. Changing classification halfway through the year without documenting the change can make month-over-month comparisons misleading.

What to Do After You Start Tracking AI Traffic

Once AI referral tracking is working, use the data to identify which content attracts useful AI-driven discovery.

Look at your highest-performing AI landing pages and ask:

What question does the page answer?

Is it informational, comparison-led or commercial?

Which AI platforms send traffic to it?

Does that traffic generate meaningful engagement?

Does it contribute to key events or pipeline?

Then use those insights to strengthen your content.

Clear answer-first passages, factual specificity, strong internal linking and useful topic coverage can make content easier for both search engines and AI systems to understand and retrieve.

Our guide to SEO content optimization explains how to improve pages for traditional search and AI citation opportunities.

For content specifically designed to be easy for AI systems to retrieve and understand, see our guide to Content for AI.

Frequently Asked Questions About AI Traffic in GA4

Start Tracking AI Traffic in GA4, but Measure Beyond GA4

It is now easier to track AI traffic in GA4 because recognised AI-assistant referrals can be separated from traditional acquisition channels.

Start with the native AI Assistant channel.

Then use Session source / medium and Explorations to understand individual AI platforms. Build custom channel rules only when you need additional control, and use UTMs only for links you genuinely control.

Most importantly, do not confuse AI referral traffic with total AI visibility.

A complete 2026 measurement framework combines GA4 referral data, AI citations and mentions, organic search performance, key events and CRM pipeline outcomes.

The goal is not simply to prove that ChatGPT, Gemini, Perplexity or Claude sent visitors.

The goal is to understand whether AI-driven discovery is helping the right buyers find your business and ultimately contributing to pipeline.