How to Track AI Referral Traffic from ChatGPT and Perplexity
Highlights
- AI assistants can send real visitors to your website through cited or recommended links.
- GA4 now has an AI Assistant channel for recognized AI referrals.
- ChatGPT search referral URLs can include utm_source=chatgpt.com.
- Perplexity traffic can be identified using its referral source and included in a custom AI channel.
- Not every AI-generated visit will be perfectly attributable.
- AI traffic should be evaluated by engagement, leads, conversions and revenue, not just sessions.
How to Track AI Referral Traffic from ChatGPT and Perplexity
Imagine you publish a detailed article about Google Ads.
A few weeks later, someone asks ChatGPT a question related to your topic. ChatGPT recommends your article as a useful source. The person clicks the link, reads your article and eventually signs up for your newsletter.
From your perspective, this is a valuable visitor.
But there is a problem.
How do you know that visitor came from AI rather than Google, a bookmark or another referral?
That is where AI referral traffic measurement becomes important.
AI platforms such as ChatGPT and Perplexity are becoming another way people discover websites. For SEO beginners, the important point is simple: ranking in traditional search is no longer the only discovery path worth measuring.
What Is AI Referral Traffic?
AI referral traffic is website traffic generated when a visitor clicks a link from an AI assistant or AI-powered search experience.
For example:
- A user asks ChatGPT a question.
- ChatGPT provides an answer and cites your website.
- The user clicks your link.
- The user arrives on your website.
- Your analytics platform records the visit.
That visit is AI referral traffic.
The same basic process can happen with Perplexity and other AI platforms.
This is different from AI visibility.
Your website might be mentioned by an AI system without anyone clicking your link. That creates visibility but not a website session.
Therefore:
AI visibility ≠ AI referral traffic.
Referral traffic measures the visitors who actually reach your website.
Why Should You Measure ChatGPT and Perplexity Traffic?
For years, website owners mainly watched Google organic traffic.
You might have checked:
- Organic sessions
- Clicks
- Impressions
- Rankings
- Conversions
Now there is another question worth asking:
Are AI assistants sending people to my website?
This matters because a visitor arriving after asking an AI assistant a specific question may already have a strong information need.
Instead of simply counting visitors, look at what those visitors do.
For example:
| Metric | What it tells you |
| Sessions | How much AI traffic you receive |
| Engaged sessions | Whether visitors interact with your site |
| Landing pages | Which content attracts AI visitors |
| Key events | Whether visitors take important actions |
| Leads | Whether AI traffic creates prospects |
| Revenue | Whether AI traffic contributes commercially |
This turns AI traffic from an interesting number into a measurable marketing channel.
How ChatGPT Referral Traffic Can Be Measured
There is good news for publishers.
OpenAI’s current publisher guidance says websites that allow OAI-SearchBot to access their content can track referral traffic from ChatGPT using analytics platforms such as Google Analytics. OpenAI also states that ChatGPT search referral URLs automatically include the UTM parameter utm_source=chatgpt.com. (OpenAI Help Center)
That gives website owners a useful signal.
In practical terms, you may see traffic associated with:
chatgpt.com
or traffic containing:
utm_source=chatgpt.com
The exact reporting appearance can depend on your analytics configuration and the particular ChatGPT experience involved.
Example
Suppose your article is:
/google-ads-conversion-tracking/
A user discovers it through ChatGPT and clicks through.
Your analytics may record information associated with ChatGPT as the traffic source.
You can then ask:
- How many users arrived?
- Which page did they enter?
- How long did they engage?
- Did they visit another page?
- Did they subscribe?
- Did they generate a lead?
That is much more useful than simply knowing that ChatGPT mentioned your website.
How to Find AI Traffic in Google Analytics 4
Google made an important change in 2026.
On May 13, 2026, Google Analytics introduced an AI Assistant channel in its default channel-group reporting. Google says recognized AI-assistant referrals can receive the ai-assistant medium and be categorized under the AI Assistant channel. (Google Support)
So your first step should be checking GA4 rather than immediately installing another tracking system.
Step 1: Open Google Analytics
Go to your GA4 property.
Then open:
Reports → Acquisition → Traffic acquisition
Look at your traffic-channel information.
Depending on your property’s data and traffic sources, you may see AI Assistant.
Step 2: Check the Traffic Source
Use dimensions such as:
Session source / medium
Look for AI-related sources such as:
- chatgpt.com
- perplexity.ai
- other recognized AI platforms
The exact sources shown depend on the traffic your website has actually received.
Step 3: Check Your Landing Pages
This is an important step that beginners often skip.
Don’t just ask:
“How many ChatGPT visitors did I receive?”
Ask:
“Which pages are ChatGPT visitors landing on?”
You might discover that three articles generate most of your AI referral traffic.
Those articles deserve closer attention.
Study their:
- Topics
- Structure
- Original research
- Definitions
- Examples
- Internal links
- Author information
- Citations and references
You can then use those insights when improving other content.
How to Measure Perplexity Referral Traffic
Perplexity works slightly differently from traditional search because its answers commonly present sources that users can open.
When someone clicks through to your website, your analytics may identify perplexity.ai as the referral source.
In GA4, check:
Reports → Acquisition → Traffic acquisition
Then inspect:
Session source / medium
Search for:
perplexity
If Perplexity traffic is present, you can analyze it separately.
What should you measure?
Suppose you discover:
- Perplexity → Article A
- Perplexity → Article B
- Perplexity → Article C
Don’t stop there.
Compare these pages for:
- Engagement
- Key events
- Leads
- Revenue
- Returning visitors
This tells you whether Perplexity is merely producing visits or producing useful business outcomes.
What If GA4 Shows Very Little AI Traffic?
This is where AI traffic measurement becomes tricky.
You may think:
“My GA4 report shows almost no ChatGPT traffic, so AI isn’t sending visitors.”
That conclusion can be wrong.
Referral attribution depends on information being passed along with the visit. If the referrer or campaign information isn’t available, analytics may not be able to confidently identify the original AI source.
Some AI-generated visits can therefore appear under broader categories such as Direct or otherwise fail to provide a clean AI attribution.
Google’s own documentation recommends custom channel groups as another way to organize AI-assistant traffic. It specifically provides an example using sources including ChatGPT, Gemini, Copilot, Claude and Perplexity. (Google Support)
So treat visible AI referrals as measured traffic, not necessarily a perfect measurement of every AI-influenced visit.
Creating an AI Traffic Channel in GA4
If you want more control, create a custom channel group.
Google’s recommended approach is to create an AI Assistants channel and use a condition such as matches regex to identify relevant AI URLs. Google also recommends placing this channel above the general Referral channel so the traffic is classified correctly. (Google Support)
For a beginner, the concept is more important than memorizing a complicated regex.
Your goal is simply:
AI sources → AI channel
rather than:
AI sources → mixed into Referral
You can include platforms relevant to your audience, such as ChatGPT and Perplexity.
Review the sources periodically because AI platforms and their domains can change.
Don’t Measure AI Traffic by Sessions Alone
Here’s a common mistake.
A website receives 100 AI visitors and 1,000 Google visitors.
It is tempting to conclude that Google is ten times more valuable.
But traffic volume doesn’t tell the whole story.
Imagine the AI visitors produce:
- 10 newsletter registrations
- 3 enquiries
- 2 purchases
while the much larger Google audience produces fewer meaningful actions.
The lesson isn’t that AI is automatically better.
The lesson is:
Measure outcomes, not just traffic.
For each AI source, examine:
Traffic → Engagement → Key events → Leads → Revenue
That gives you a much clearer picture of value.
AI Referral Traffic vs AI Visibility
These two concepts should remain separate in your SEO reporting.
AI visibility
Your content is:
- Mentioned
- Cited
- Recommended
- Included as a source
But the user may never visit your website.
AI referral traffic
The user actually clicks your link and arrives on your website.
Therefore, a page can have:
High AI visibility + low referral traffic
or:
Low visible AI traffic + strong business results from a smaller number of visitors.
Both situations require different strategies.
How to Improve Your AI Referral Measurement
-
Track AI sources separately
Don’t throw every AI platform into one unexplained number.
Where possible, distinguish:
- ChatGPT
- Perplexity
- Gemini
- Claude
- Copilot
- Other relevant AI sources
-
Monitor landing pages
Find the pages receiving AI referrals.
These pages provide clues about the subjects AI systems consider useful for users.
-
Track conversions
Configure important events such as:
- Newsletter signup
- Contact form submission
- Product purchase
- Demo request
- Download
- Registration
-
Compare AI with organic search
AI traffic should complement your existing SEO reporting.
Compare AI visitors with:
Organic Search vs AI Assistant vs Referral vs Direct
This helps you understand how users discover your content across different channels.
-
Don’t assume missing attribution means missing traffic
If AI traffic appears unusually low, investigate your analytics setup before concluding that AI platforms aren’t sending visitors.
A Simple AI Traffic Measurement Example
Imagine you run an SEO blog.
During one month, your analytics shows visitors arriving from ChatGPT and Perplexity.
You discover that most AI visitors land on:
“How to Improve Website Speed”
You then notice those visitors are also reading:
“What Is Core Web Vitals?”
That gives you an interesting content journey:
AI answer → Speed article → Core Web Vitals article → Newsletter signup
Now you have something actionable.
You can strengthen the internal links between those articles and create additional content around closely related questions.
This is how AI referral measurement can become part of an actual SEO strategy rather than simply another dashboard number.
Key Takeaways
- AI referral traffic is traffic generated when users click through from AI platforms to your website.
- ChatGPT and Perplexity can send visitors through links and citations.
- Google Analytics 4 now provides an AI Assistant channel for recognized AI referrals. (Google Support)
- OpenAI says ChatGPT search referral URLs automatically include utm_source=chatgpt.com. (OpenAI Help Center)
- Perplexity traffic can be investigated through referral-source information and custom GA4 channel grouping.
- Don’t assume analytics captures every AI-influenced visit perfectly.
- Measure engagement, conversions, leads and revenue, not sessions alone.
FAQ
-
What is AI referral traffic?
AI referral traffic is website traffic that arrives after a user clicks a link from an AI assistant such as ChatGPT or Perplexity.
-
Can I track ChatGPT traffic in Google Analytics?
Yes. OpenAI states that publishers can use analytics platforms such as Google Analytics to track ChatGPT referral traffic, and ChatGPT search referral URLs include utm_source=chatgpt.com. (OpenAI Help Center)
-
How can I track Perplexity traffic?
In GA4, check Acquisition → Traffic acquisition and examine Session source / medium for Perplexity-related referral traffic. You can also create a custom AI-assistant channel to organize it separately.
-
Does AI referral traffic include every person who sees my website mentioned by an AI?
No. Referral traffic measures visits to your website. Someone can see your website mentioned or cited by an AI system without clicking the link.
-
Why is my AI traffic lower than expected?
Some AI-influenced visits may not carry enough attribution information for analytics to identify their original source. Therefore, visible AI referral traffic should not automatically be treated as a complete count of all AI-driven discovery.
Conclusion
AI is creating another route between your content and your audience.
For SEO beginners, the important change is not that you need to abandon Google SEO. Instead, you should start asking a broader question:
How are people discovering my content across search engines and AI assistants?
ChatGPT and Perplexity are two important sources to monitor. With GA4’s newer AI Assistant reporting, ChatGPT’s UTM tracking, and custom channel grouping for additional platforms, you can build a much clearer picture of AI referral traffic. (Google Support)
The smartest approach is to go beyond counting AI visitors. Find the pages they visit, understand what they do, and measure whether those visits create meaningful results.
AI visibility gets attention. AI referral measurement tells you what happens after the click.