How to Track AI Traffic in GA4 Using Custom Channel Groups

AI traffic is real. It is showing up in GA4 already. Most businesses just are not tracking it properly.

If people are clicking through from ChatGPT, Gemini, Perplexity, Copilot, or Claude, that traffic often gets dumped into Referral by default. Which means it is there, but buried. And if it is buried, it is hard to measure, hard to report on, and easy to ignore.

The good news is you do not need another tool to fix that.

You can track AI traffic in GA4 by creating a custom channel group that pulls those visits into their own bucket. It is simple, cheap, and a lot more useful than leaving everything shoved under Referral.

If you want the short version: that is the answer.

If you want the step-by-step, regex examples, reporting tips, and the bits most articles miss, keep reading.

Can you track AI traffic in GA4?

Yes.

GA4 can track traffic from AI tools when the visit includes a detectable source or referrer. In plain English, that means if someone clicks through to your site from something like ChatGPT or Perplexity, GA4 can often see where they came from.

The issue is not whether GA4 can see it.

The issue is that GA4 usually throws it into Referral, which is technically correct but not very helpful.

A click from ChatGPT is not the same as a click from some random directory, vendor site, or blog backlink. Treating them all the same gives you weak reporting and weaker decisions.

Key takeaway: GA4 can track AI traffic, but you usually need a custom channel group to make that data useful.

Why AI traffic usually shows up as Referral

GA4 uses channel grouping rules to decide where traffic belongs.

Its default channel grouping is built by Google and cannot be directly edited. So if GA4 sees a visit from an external site and it does not match a more specific channel rule, it often ends up under Referral.

That is why ChatGPT traffic analytics in GA4 can be awkward out of the box. The traffic is often there, but it is mixed in with everything else.

This creates a few problems:

you cannot easily report on AI traffic as its own channel
you cannot compare AI traffic properly against Organic Search, Paid Social, Email, or Direct
you cannot clearly see whether AI discovery is growing over time
you miss the chance to connect AI visibility with leads, sales, or engagement

This is why so many people are now looking for ways to break out LLM traffic in GA4 properly.

Should you separate AI traffic from Referral?

Separate AI traffic

Cleaner acquisition reporting by channel.
Easier comparisons with Organic Search, Paid, Social, and Direct.
Clearer trend tracking for AI discovery over time.
More useful reporting for teams and clients.

Leave it inside Referral

Reporting stays basic but less informative.
AI visits remain mixed with unrelated referral sources.
Harder to connect AI discovery to leads, sales, or engagement.
Usually only acceptable when AI traffic volume is near zero.

Why custom channel groups are the best fix

There are a few ways you could look at AI traffic in GA4.

You could use filters. You could use explorations. You could build custom reports.

But for most businesses, the best option is a custom channel group.

Why? Because it gives you a reusable reporting structure instead of a one-off workaround.

That means:

cleaner acquisition reporting
easier comparison against other channels
less faffing about every time you want to check performance
better reporting for clients, internal teams, or anyone who does not live in GA4 all day

It is also a much more sensible approach than buying another tool before you have even sorted your GA4 basics.

Key takeaway: a custom channel group for AI traffic is usually the simplest and most practical setup.

Which AI sources should you include?

At a minimum, most businesses should include:

ChatGPT / OpenAI
Google Gemini
Microsoft Copilot
Claude / Anthropic
Perplexity

You can expand the list later if new sources show up in your data.

There is no prize for building the world's most bloated regex on day one. Start with the obvious platforms. Then review your source data and add more if needed.

That is the smarter approach.

How to track AI traffic in GA4 step by step

GA4 setup flow

1
Step 1: Go to Channel groups in GA4
In GA4: Admin -> Channel groups.
2
Step 2: Create a custom channel group
Create new or duplicate existing; do not try to edit the default channel group.
3
Step 3: Add a new channel called AI Traffic
Name it clearly so non-technical stakeholders understand it quickly.
4
Step 4: Add Source matches regex rule
Match key AI assistant sources such as ChatGPT, Claude, Gemini, Copilot, and Perplexity.
5
Step 5: Move AI Traffic above Referral
Ordering matters or GA4 may classify visits as Referral first.
6
Step 6: Save and test in acquisition reports
Apply the custom channel group in Traffic Acquisition and validate classification.

Step 1: Go to Channel Groups in GA4

In GA4, head to:

Admin -> Channel groups

This is where you create and manage custom channel definitions.

Step 2: Create a custom channel group

Do not waste time trying to edit the default channel group. You cannot.

Instead, create a new custom channel group or duplicate an existing one.

Call it something obvious like:

Default + AI Traffic
AI Traffic Channel Group
Qwestyon Custom Channels

Keep it simple. You want future-you to know exactly what it is.

Step 3: Add a new channel called "AI Traffic"

Create a new channel inside your custom group.

You can call it:

AI Traffic
AI Assistants
LLM Traffic

We would usually go with AI Traffic because it is clearer for normal humans. "LLM traffic" is accurate, but a lot of business owners do not speak in model terminology, and they should not have to.

Step 4: Set the rule to match AI sources

This is the important bit.

Use a rule based on Source matches regex.

A strong starter regex is:

(chatgpt|openai|claude|anthropic|gemini|copilot|perplexity)\.(com|ai)

That should catch the main AI assistant sources most businesses care about.

If you want a broader version, use:

.*openai.*|.*chatgpt.*|.*claude.*|.*anthropic.*|.*gemini.*|.*copilot.*|.*perplexity.*

The first regex is cleaner. The second is broader and more forgiving.

Start simple unless your source data tells you otherwise.

Step 5: Move AI Traffic above Referral

This part is easy to miss and important to get right.

Make sure your AI Traffic channel sits above Referral in the rule order.

If it sits below Referral, GA4 may classify those visits as Referral first, and your AI channel will not catch them properly.

That would be a very annoying own goal.

Step 6: Save and test the setup

Once the channel group is saved, head to:

Reports -> Acquisition -> Traffic acquisition

Then use your custom channel group in the report.

You should start to see AI Traffic appear as its own channel if GA4 is detecting those visits.

You can also use the same setup in User Acquisition, Explorations, and Looker Studio.

Regex examples you can copy

Here are a few practical regex options for tracking AI referral traffic in GA4.

Basic regex

Use this if you want something clean and readable:

(chatgpt|openai|claude|anthropic|gemini|copilot|perplexity)\.(com|ai)

Broader regex

Use this if you want to catch more variations:

.*openai.*|.*chatgpt.*|.*claude.*|.*anthropic.*|.*gemini.*|.*copilot.*|.*perplexity.*

Keep it maintainable

Do not build some cursed mega-regex just because someone on LinkedIn wanted to sound clever.

If your regex is too messy to understand or update, it becomes a pain to maintain. Start with the major sources. Add more when your own GA4 source data gives you a reason.

That is a much better way to work.

When should you separate AI traffic from Referral?

Usually, you should separate it.

It makes sense to break AI traffic out into its own channel if:

you care about AI search visibility or GEO
you want clearer acquisition reporting
you want to compare AI traffic against Organic Search or Social
you want to see whether AI visits actually convert
you report on content performance and landing pages

You might leave it inside Referral if:

your reporting is extremely basic
you get virtually no AI traffic yet
nobody is going to use the segmented data anyway

But for most sites, this is worth doing now rather than later.

The setup is quick. The upside is clear. The downside is basically nonexistent.

Key takeaway: if AI traffic matters even slightly to your business, separating it from Referral usually makes reporting much more useful.

What to look at once the data starts coming in

This is where the good stuff starts.

A lot of articles show you the setup and then stop dead. But the whole point of tracking AI traffic is using the data.

Here is what to check first.

Landing pages

Which pages are getting AI visits?

This helps you spot the content that AI tools are surfacing and citing most often. Those pages are strong candidates for:

better internal linking
stronger CTAs
clearer intros
fresher examples
deeper supporting content

If AI tools keep sending people to a page, that page is doing something right.

Engagement rate

Are AI visitors actually engaging, or bouncing straight off?

If engagement is poor, that might mean your page got picked up for a narrow answer but does not really satisfy the wider intent once people land.

That is useful to know.

Key events and conversions

This matters more than vanity traffic.

Track whether AI visitors:

submit forms
book calls
sign up
buy
engage with important actions

Traffic is nice. Commercial value is nicer.

Source-level breakdown

Once your AI channel is working, break it down by source.

That helps you see whether the visits are mainly coming from:

ChatGPT
Perplexity
Gemini
Copilot
Claude

That is useful because not all AI traffic behaves the same way.

Comparison with Organic Search

This is where things get properly interesting.

Compare AI traffic with Organic Search on:

engagement
landing pages
conversion rate
assisted conversions
user quality

That gives you a much better sense of whether AI discovery is just a curiosity or an actual growth channel.

A simple Looker Studio view for AI traffic

You do not need to build some ridiculous dashboard monster here.

A simple Looker Studio page is enough.

Include:

scorecards for Users, Sessions, Engaged Sessions, Key Events, Leads, or Revenue
time series chart for Sessions over time
table with Session Source, Landing Page, Sessions, Engagement Rate, and Key Events
comparison filter for AI Traffic vs Organic Search vs Referral

That gives you a clean view of:

whether AI traffic is growing
which platforms are sending it
which pages are attracting it
whether it is commercially useful

That is enough to start learning from the data without drowning in dashboard fluff.

Important limitations and gotchas

Not all AI influence becomes trackable traffic.

This is worth saying clearly.

Not every AI-driven visit shows up neatly in GA4.

Someone might:

read about your brand in ChatGPT and come back later via Google
copy and paste your URL directly
mention you in a meeting and someone else visits later
find you through an AI summary but never click the cited source

So GA4 can show you part of the picture, but not all of it.

That does not make this setup pointless. It just means you should avoid acting like GA4 is a complete AI visibility tracker. It is not.

Your regex may need updating.

AI platforms change. Domains change. New tools pop up.

So review your source data every so often and update your regex if needed.

This is not difficult. It is just basic maintenance.

Channel group order matters.

Worth repeating because it catches people out:

Your AI Traffic channel needs to sit above Referral.

Otherwise you may accidentally lose the clean classification you were trying to create in the first place.

Small traffic volumes are normal at first.

Do not panic if the numbers look tiny.

For a lot of businesses, AI traffic is still early-stage. Small numbers do not mean the setup is wrong. They often just mean the channel is still emerging.

That is fine.

The important thing is that you are measuring it now instead of waking up six months late.

Quick diagnostic checklist

✓Is your AI Traffic custom channel above Referral in rule order?
✓Are your regex rules simple, readable, and based on current source data?
✓Are you validating results in Traffic Acquisition with the custom channel group selected?
✓Are you comparing AI Traffic against Organic Search and Referral over time?
✓Are you reviewing landing pages and conversion quality, not just sessions?
✓Are you reviewing source data periodically and updating regex as platforms evolve?

Should you track AI traffic separately from referral traffic?

Yes, in most cases.

If your goal is better reporting, better content insight, and a clearer view of how AI tools are influencing discovery, then separating AI traffic from generic referral traffic is the sensible move.

It gives you cleaner data without adding extra software, extra cost, or extra complexity.

And frankly, that is a rare win.

Final takeaway

If you want to know how to track AI traffic in GA4, here is the no-BS answer:

Create a custom channel group, add an AI Traffic channel, use a source regex for platforms like ChatGPT, Gemini, Claude, Copilot, and Perplexity, and place that channel above Referral.

That is the simplest way to track LLM traffic in GA4 and make the data actually useful.

No extra tool. No overcomplicated workaround. No pretending Referral is "good enough."

Just cleaner reporting and a better read on how people are finding you through AI.

Suggested Internal Resources

What Is Cross-Network Traffic in GA4?
Why is my Meta Traffic Showing as Unassigned in GA4?
GEO service page
Google Ads and analytics support
A post on how to measure content performance in GA4
A post on Looker Studio dashboards for small businesses

FAQ

AI Traffic in GA4 FAQs

How do I track ChatGPT traffic in GA4?

Create a custom channel group in GA4 and add a source regex that matches ChatGPT or OpenAI traffic. Then place that AI Traffic channel above Referral.

Does GA4 automatically track AI traffic?

GA4 can detect some AI traffic when source or referrer data is available, but it usually classifies it as Referral by default instead of as a separate AI channel.

What is a good regex for AI traffic in GA4?

Should AI traffic be grouped with Referral?

It can be, but in most cases it is more useful to separate it so you can compare AI traffic against Organic Search, Paid, Social, and other channels.

Can I track LLM traffic in GA4 without another tool?

Yes. For most businesses, custom channel groups in GA4 are the simplest and cheapest way to do this.

Can I report on AI traffic in Looker Studio?

Yes. Once your custom channel group is set up in GA4, you can use it in Looker Studio to report on sessions, sources, landing pages, engagement, and conversions.