#Why Tracking AI Tool Traffic Matters
AI tools like ChatGPT, Perplexity and Claude are increasingly shaping how people discover content. Someone asks a question, the tool generates an answer, and if your page is cited the reader clicks through.
Knowing this helps you:
- Understand which of your content AI tools are recommending
- Measure engagement from AI-sourced visitors
- Double down on the content that's already getting traction, and roll those learnings out across other pages
These visitors also tend to arrive with intent already formed. They've had part of the conversation inside the AI tool before they land on you, which often makes them closer to converting than cold search traffic.
#How AI Tools Refer Traffic
Each platform passes referral data differently, which is exactly why this is fiddly. The lay of the land:
- ChatGPT (OpenAI): Passes a referrer (chatgpt.com) on the desktop web version in many cases. But links from the mobile app, and some inline links, strip the referrer, so those visits land in Direct.
- Perplexity AI: Usually passes a visible referrer, typically perplexity.ai, especially when users follow citations.
- Claude (Anthropic): Passes a referrer (claude.ai) on desktop web. Like the others, app traffic is inconsistent. Claude is now a meaningful source in its own right, particularly for B2B content, so it's no longer one to write off.
- Microsoft Copilot: Frequently shows up via copilot.microsoft.com or bing.com.
- Gemini (Google): Appears via gemini.google.com when accessed in a browser, though it can be inconsistent.
The pattern across all of them: if the referrer is passed, GA4 can see it. If it isn't, the visit falls into Direct and there's no signal left to recover it.
#The Big Change: GA4 Now Has a Native AI Assistant Channel
This is the part that's new, and it's the reason this guide looks different from the version you may have read last year.
In May 2026, Google added an AI Assistant channel to GA4's Default Channel Group. It requires no setup. When a session arrives with a referrer that matches a recognised AI tool, GA4 tags it with the medium ai-assistant and files it under AI Assistant automatically. Rollout was gradual, so if you didn't see it straight away that's normal; broad availability landed in early June 2026.
At launch Google named three recognised sources: ChatGPT, Gemini and Claude. The list is expected to grow, and reporting on which tools are covered has been inconsistent, so check your own account rather than taking a blog's word for it (this one included).
To find it: go to Reports → Acquisition → Traffic acquisition, then set the primary dimension to Session default channel group. If you've had qualifying AI traffic since the channel went live, “AI Assistant” shows up as its own row.
Two things to know before you trust the number:
- It only counts forward. GA4 processes data forward only, so AI visits from before the channel launched stay in their old buckets (Referral or Direct). That means no back history and no clean year-on-year comparison through the native channel yet.
- It only sees referrers. If a tool strips the referrer, the visit never reaches this channel. Treat the AI Assistant number as a floor, not a ceiling.
#Why You Still Want a Custom Channel Group
The native channel is a great head start, but it has two gaps a custom channel group fills.
First, history. Unlike the native channel, a custom channel group reclassifies your existing data when you save it, so you can look back over previous months rather than starting from zero.
Second, coverage. The native channel sticks to a short list of recognised tools. A custom group lets you capture anything passing a referrer, including newer or less common sources, and keep adding to it as new tools appear.
So the ideal setup is both: the native channel for the easy ongoing read, and a custom group for your back catalogue and the long tail.
#Step-by-Step: Build a Custom AI Channel Group in GA4
#1. Open GA4 → Admin
- Go to your GA4 property
- Under Property, click Data Settings → Channel Groups
#2. Create a New Channel Group
- Click Create new channel group
- Name it something clear like AI Tools Traffic
- Click Add new channel
#3. Add a Regex Rule for Known AI Tools
- Under the new channel, click Add condition
- Choose Session source
- Select matches regex
- Paste this single anchored pattern:
^(chatgpt\.com|chat\.openai\.com|openai\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|bard\.google\.com|copilot\.microsoft\.com|deepseek\.com|grok\.com|x\.ai|meta\.ai|you\.com|phind\.com|mistral\.ai)$One clean, anchored list beats the old habit of stacking several overlapping rules. Avoid loose patterns like .*ai.*, which match any source with the letters “ai” in it (think email providers, retail domains, plenty of others) and quietly inflate the channel.
#4. Order the Rule Above Referral
This is the step most people miss. GA4 reads channel rules top to bottom. If an AI source hits the Referral rule before it reaches your AI rule, it gets filed as Referral instead. Drag your AI channel above Referral in the group so it's matched first. Rule order, not the regex, is the usual reason a freshly built group still shows AI traffic under Referral.
#5. Label and Save
- Name the channel AI Tools
- Set the channel type to Custom
- Click Save to add it, then Publish to make it available in reports
#6. Use in Reports
- Go to Reports → Acquisition → Traffic acquisition
- Switch the channel group selector at the top of the table to your new group
- You'll see your AI Tools row with session data
Add Landing page as a secondary dimension here. Seeing which of your pages AI tools send people to is the single most useful thing this whole exercise tells you.
#A Word on AI Overviews
Worth clearing up, because people conflate the two constantly: Google's own AI Overviews and AI Mode clicks are counted as Organic Search in GA4, not as AI traffic. AI Mode links also launched without referrer data, which pushes some of those clicks into Direct. So the AI Assistant channel is about AI assistants sending you visitors, not about Google's AI features inside search. Keep them mentally separate when you report.
#Pro Tips: Making AI Traffic Easier to Track
- Tag links you share in AI-adjacent spaces with UTMs (for example
?utm_source=chatgpt&utm_medium=ai). Even when some tools strip parameters, the ones that pass through give you clean attribution. - Watch your Direct trend. A rising Direct line alongside flat branded search is often AI traffic with no referrer. You can't itemise it, but you can acknowledge it.
- Review the regex quarterly. New tools and new domains appear regularly. Skim your Referral source list each month and add anything new you spot.
#Limitations and What to Watch Out For
- Most AI tools still don't pass a referrer consistently, and app traffic almost never does
- A large share of real AI visits - often well over a third - sits permanently in Direct, where no rule can reach it
- The native channel has no back history, so build the custom group if you want to look backwards
- Don't read a single month as a trend. Annotate the May 2026 launch in your reports and give it a 90-day baseline before drawing conclusions
#FAQs
Can I track ChatGPT traffic in GA4?
Yes, when ChatGPT passes a referrer (common on desktop web). App traffic and some inline links strip the referrer, so a portion still lands in Direct. The native AI Assistant channel picks up the referred visits automatically.
Does Perplexity AI appear in GA4?
Yes, it typically shows up as perplexity.ai. Whether it falls inside the native AI Assistant channel depends on Google's recognised-source list, so confirm in your own reports and keep it in your custom regex either way.
Can I track Claude traffic?
Yes. Claude is one of the three sources Google named at launch, and claude.ai passes a referrer on desktop web. It's become a genuine source for B2B content, so it's worth watching.
Do I still need a custom channel group now that GA4 has a native one?
For most people, yes. The native channel only counts forward and sticks to a short list of tools. A custom group gives you your back history and broader coverage.
Is it worth tracking AI referrals at all?
Absolutely. The volume is still modest for many sites, but it's growing fast and these visitors tend to be high-intent. Getting the measurement in place now beats scrambling once the numbers are too big to ignore.
AI discovery is growing fast, and some of that traffic is quietly arriving at your site right now. The good news is you no longer have to build everything from scratch: switch on the native AI Assistant channel for an instant read, then add a custom channel group for your history and the sources it misses.
Do both, keep an eye on your Direct bucket, and you'll have an honest picture of what AI is sending your way, which content it's recommending, and where to double down.
This guide is written by Hedgehog, a DIY digital marketing consultancy specialising in small and medium businesses in Australia. We offer digital marketing consulting, coaching and training.
