AI traffic in GA4, how to see visits from ChatGPT, Perplexity and other AI tools

Can GA4 show traffic from AI search?

Yes. Visits from AI tools arrive with a referrer, such as chatgpt.com or perplexity.ai, and GA4 records it as the session source. That is what AI referral traffic means in practice.

GA4 also has a built-in answer now. Google's default channel group includes an AI Assistant channel. Google describes it as the channel for users who arrive from sources like ChatGPT, Gemini, Deepseek, Copilot or Grok. A visit counts as AI Assistant when the medium is ai-assistant or the referrer is on Google's list of AI assistants.

That is a good start. But there are reasons to build your own view as well, and this guide walks through both.

What does the AI Assistant channel leave out?

It leaves out Google's own AI features. Google says the channel excludes AI Overviews and AI Mode. Clicks from those arrive as Google organic search, because they happen on a Google results page.

It also depends on Google's list. A newer or smaller AI tool may not be on it yet, so its visits land in Referral. And you cannot see the list or change it.

Finally, it only groups the traffic. To see which AI tool sent a visit, which page it landed on and what the visitor did next, you need an exploration. The steps below cover that.

How do you check the AI Assistant channel?

Open the standard traffic report and look for the row. It takes a minute.

1. In GA4, go to Reports, then Acquisition, then Traffic acquisition.
2. Make sure the table's first column is Session default channel group.
3. Look for AI Assistant in the list, and set a date range of at least three months.
4. Click the plus next to the column name and add Session source as a secondary dimension.

You will now see AI Assistant visits split by source. For most sites, chatgpt.com is the largest, but it varies. If there is no AI Assistant row, you may simply have had no visits Google recognised in that period.

How do you build an AI traffic exploration in GA4?

An exploration gives you a flexible table you can filter with a regex. Google's explorations help page covers the basics. Here is the setup for AI traffic.

1. Click Explore in the left menu, then choose Blank.
2. Name it something clear, such as AI referral traffic.
3. In the Variables column, click the plus next to Dimensions. Import Session source, Session medium, Landing page plus query string and Session default channel group.
4. Click the plus next to Metrics. Import Sessions, Engaged sessions, Key events and Total revenue if you sell online.
5. Drag Session source into Rows, then Landing page plus query string under it.
6. Drag your metrics into Values.
7. Under Filters, choose Session source, then matches regex, and paste the pattern below.

This is a reasonable starting pattern.

.*(chatgpt|openai|perplexity|gemini\.google|bard\.google|claude\.ai|copilot|deepseek|grok|meta\.ai|mistral|you\.com|poe\.com).*

GA4 uses full match regex by default, as Google's regex help page explains. The .* at each end lets the pattern match anywhere in the source value. The backslash before each dot makes it a real dot, not "any character". GA4 uses Google's RE2 syntax, so avoid features RE2 does not support, such as lookaheads.

Set the date range to the last 90 days or more and save. You now have AI visits by tool and by landing page, with engagement and key events alongside.

How do you create a custom AI channel group?

A custom channel group puts your own AI channel into the standard reports. Google's channel groups help page lists the steps. Here they are for an AI channel.

1. Go to Admin, then Data display, then Channel groups.
2. Find the default channel group and choose to copy it, so all the normal channels come across.
3. Name the new group, for example Channels with AI.
4. Click Add new channel and call it AI search and chat.
5. Add a condition on Source, choose matches regex, and paste the same pattern.
6. Save the channel, then click Reorder and drag it near the top, above Referral.
7. Click Save group.

The order step is the one people miss. Google says traffic goes into the first channel whose definition it matches. If Referral sits above your AI channel, Referral catches the visits first and your AI channel stays empty.

You can now pick this group in the traffic reports. Google says custom channel groups apply to your reports retroactively, so older AI visits show up too. A standard property allows two custom channel groups, so use them wisely.

How do you check the regex is catching everything?

Look at what it misses. Once a quarter, open an exploration with no filter, put Session source in Rows, and sort by sessions. Scan the list for any AI tool your pattern does not cover.

New tools appear often, and existing ones change their domains. When you spot one, add it to the pattern in both the exploration and the channel group. Keep a note of the date, so anyone reading the report knows when the definition changed.

It is also worth checking the other way. A loose word in the pattern, such as copilot, may catch a source that is not an AI search tool at all. If a strange source appears in your AI channel, tighten the pattern.

Should you use the default channel or your own?

Use both, for different jobs. The default AI Assistant channel is fine for a quick monthly figure. Your own channel group and exploration are better for analysis and reporting.

If the totals differ between your channel and Google's, check the Referral channel for AI sources that Google does not list yet. Then decide whether to add them to your regex.

Why is AI traffic in GA4 always an undercount?

Because some AI visits never show an AI source. There are three common reasons.

Missing referrers. Some apps do not pass a referrer when a user taps a link. People also copy a link from an answer and paste it into a browser. Those visits land in Direct.

Google AI features. Clicks from AI Overviews and AI Mode count as Google organic. Google's guide to generative AI features points to Search Console's Generative AI performance report for measuring those instead.

Influence without a click. Many people read an AI answer, remember a brand name and search for it later. That shows up as branded search or direct traffic, not AI traffic.

So treat GA4's AI figures as the floor, not the full picture. Pair them with visibility tracking to see how often you are named, whether or not anyone clicks. The guide on how to measure AI visibility covers that side.

How do UTM tags help with ChatGPT traffic?

They make ChatGPT visits easy to identify. OpenAI's publisher FAQ says ChatGPT automatically adds utm_source=chatgpt.com to referral URLs.

That means ChatGPT search visits can show a source of chatgpt.com even when the referrer is lost. It also means you will see that parameter in landing page reports. If your site strips query strings on redirect, check that the tag survives, or you lose that signal.

What should you report each month?

Keep it short and consistent. A useful monthly view has a handful of numbers.

• AI sessions, split by source.
• Engaged sessions and engagement rate from AI, compared with organic search.
• Key events or revenue from AI visits.
• The top landing pages for AI traffic.
• The change against last month and the same month last year.

The landing pages list is often the most useful part. It shows which pages AI engines send people to, which usually matches the pages they cite. Those pages deserve the most care.

How Axiom GEO helps

AI traffic analytics connects Google Analytics 4 and Search Console with read-only access and syncs daily. It shows AI referral traffic from chatgpt.com, perplexity.ai, gemini.google.com, claude.ai, Copilot and others, and joins sessions and clicks at page level. Alongside, AI visibility tracking shows how often you are named in answers, including the ones that send no click. Client reports can include both in a white label monthly PDF.