Track AI referral traffic by creating a clean “AI assistants” traffic group, then compare it against leads, sales, and engaged sessions. GA4 can do this, but it needs more setup. Matomo is often easier if you want clearer referrer data and more control. GA4 became the main Google analytics tool after Universal Analytics stopped processing standard website data on July 1, 2023 [source: Google].
TLDR: AI traffic is traffic from tools like ChatGPT, Perplexity, Claude, Gemini, and Copilot. In GA4, you track it with custom channel groups, filters, and reports. In Matomo, you can usually inspect and group referrers with less pain. Example: if your blog gets 10,000 sessions, and 120 come from Perplexity, 45 from ChatGPT, and 25 from Copilot, that is 1.9% AI referral traffic worth watching.
What is AI referral traffic?
AI referral traffic is website traffic sent by AI tools.
A person asks a chatbot a question. The chatbot answers. It may cite your page or link to your site. The person clicks. Boom. You have an AI referral.
Common AI referrers include:
- chatgpt.com
- perplexity.ai
- claude.ai
- gemini.google.com
- copilot.microsoft.com
- you.com
- poe.com
The annoying part is simple. Not every AI visit shows up neatly. Some tools hide referrer data. Some open links inside apps. Some send traffic that looks like direct traffic. Honestly, it feels like analytics decided to wear a fake mustache.
Why AI traffic matters
AI search is not normal search.
Google search gives users a list of blue links. AI tools often give users a direct answer. Your link may appear as a citation, a footnote, or a source card. That changes user behavior.
AI visitors may also be very warm. They asked a specific question. They got your page as part of the answer. Then they clicked.
That means AI traffic can be small but valuable.
For example:
- You get only 80 AI sessions in a month.
- Those users spend 2 minutes 40 seconds on site.
- They convert at 4.5%.
- Your normal referral traffic converts at 1.2%.
That is not noise. That is a signal.
How to track AI traffic in GA4
GA4 does not have a magic “AI traffic” button. Of course it does not. That would be too kind.
You need to build your own view.
Step 1: Check referral traffic
Go to:
- Reports
- Acquisition
- Traffic acquisition
Set the main dimension to Session source / medium. Then search for AI domains.
Try these:
- chatgpt
- perplexity
- claude
- copilot
- gemini
- poe
You may see rows like:
- perplexity.ai / referral
- chatgpt.com / referral
- copilot.microsoft.com / referral
Step 2: Create a custom report
GA4 standard reports are fine for a quick peek. But you will want a saved report.
Create an exploration with:
- Dimension: Session source
- Dimension: Session medium
- Metric: Sessions
- Metric: Engaged sessions
- Metric: Conversions
- Metric: Revenue, if you track it
Then filter the session source with a regex list.
Example regex:
chatgpt|openai|perplexity|claude|anthropic|gemini|copilot|you\.com|poe\.com
This catches many common AI sources. Keep adding new ones. AI tools pop up like mushrooms after rain.
Step 3: Build a custom channel group
This is the cleanest GA4 method.
Go to Admin, then Data display, then Channel groups. Create a new group called AI Assistants.
Add rules like:
- Source contains chatgpt
- Source contains perplexity
- Source contains claude
- Source contains copilot
- Source contains gemini
Now your AI traffic can sit beside organic search, paid search, referral, email, and social.
It is much easier to explain in a report. Your boss does not want 19 rows of weird domains. Your boss wants one line that says AI Assistants.
How to track AI traffic in Matomo
Matomo feels more direct for this job.
You can inspect referrers, create segments, and keep more control over your data. If you self host Matomo, you also own the raw data. That helps when you want to audit traffic patterns.
Image not found in postmetaStep 1: Open referrer reports
Go to:
- Acquisition
- All Referrers
- Websites
Look for the same AI domains.
Matomo often makes this feel less buried. You see referrer sources in a plain way. There is less report wrestling.
Step 2: Create an AI segment
Create a segment where the referrer URL or referrer name contains AI domains.
Use rules like:
- Referrer contains chatgpt
- Referrer contains perplexity
- Referrer contains claude
- Referrer contains copilot
- Referrer contains gemini
Name the segment AI Referral Traffic.
Then apply it across reports. You can check pages, goals, ecommerce, time on site, and exits.
Step 3: Use goals and ecommerce tracking
Traffic alone is cute. Money is better.
Track what AI visitors do after they land.
- Do they sign up?
- Do they buy?
- Do they read more pages?
- Do they return later?
- Do they download a guide?
If Perplexity sends 60 visits and 6 leads, that is a 10% lead rate. If ChatGPT sends 200 visits and 2 leads, that is 1%. Bigger is not always better.
GA4 vs Matomo: which is better for AI traffic?
Here is the quick version.
| Feature | GA4 | Matomo |
|---|---|---|
| Setup speed | Slower | Faster |
| Custom AI grouping | Good, but fiddly | Simple with segments |
| Raw data control | Limited unless using BigQuery | Strong, especially self hosted |
| Free version | Yes | Yes, self hosted |
| Privacy control | Moderate | Strong |
| Learning curve | Higher | Lower for referrers |
GA4 is best if you already use Google Ads, Search Console, and BigQuery. It fits the Google stack. It also works well for large teams that already live inside GA4.
Matomo is best if you want clearer ownership, easier referrer checks, and privacy friendly reporting. It is also nice when you hate clicking through four menus just to answer one basic question.
What numbers should you watch?
Do not only count sessions. That can trick you.
Watch these instead:
- AI sessions: total visits from AI tools.
- Engagement rate: did they stick around?
- Landing pages: which pages AI tools recommend.
- Conversions: signups, leads, sales, calls, demos.
- Revenue: if ecommerce is active.
- Return visits: AI may start the first visit, not the final sale.
Also compare AI traffic to organic search. If your AI visitors convert better, create more answer friendly content. Add clear definitions. Add examples. Add short summaries. Cite facts. Make pages easy for humans and machines to understand.
Common tracking problems
Expect gaps. Some AI traffic will show as direct. Some will show under odd domains. Some will vanish due to browser privacy rules.
That does not mean tracking is useless. It means you need trends, not perfection.
Update your AI source list every month. Add new tools when you spot them. Check landing pages too. A sudden jump in direct traffic to one old blog post may be AI driven, even if the referrer is missing.
Best simple setup
If you use GA4, create a custom channel group called AI Assistants. Use regex in explorations. Review it monthly.
If you use Matomo, create an AI Referral Traffic segment. Apply it to goals and ecommerce reports. Check referrer URLs often.
The best tool is the one you will actually check. GA4 is powerful. Matomo is clearer. Both can measure AI traffic. Just do not leave AI visits mixed into generic referral traffic forever. That is how useful signals get lost in the junk drawer.