How to Track ChatGPT Traffic in Analytics

Q: What is the source/medium for ChatGPT traffic in GA4?

ChatGPT traffic appears in GA4 as chatgpt.com / referral for sessions from the current ChatGPT.com interface. Older referrals from before the 2024 domain migration appear as chat.openai.com / referral. Both values must be included in your custom channel group regex to capture all ChatGPT-referred sessions. Without including both source values, sessions from the legacy domain will be miscategorized in the default acquisition report.

Q: Is AI traffic counted as direct traffic in GA4 by default?

Some AI traffic appears as direct in GA4 due to referrer header suppression. When users copy a link from a ChatGPT or Perplexity response and paste it into a new browser tab, no referrer header is sent, and GA4 records the session as direct. This is called dark traffic. The proxy signals for AI dark traffic are a spike in direct sessions with a high new user ratio above 40% and very short session duration under 10 seconds, all of which are patterns inconsistent with branded direct traffic.

Q: How do I create a custom channel group for AI traffic in GA4?

In GA4, go to Admin, then Data Display, then Channel Groups, then create a new custom channel group. Add a new channel called AI Search Traffic with a condition: Session source matches regex chatgpt\.com|chat\.openai\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|edgepilot|x\.ai. Save the channel group. The new channel appears in your Traffic Acquisition report within 24-48 hours. The regex covers ChatGPT, Perplexity, Claude, Gemini, Bing Copilot via Edge, and Grok.

Q: How do I track Perplexity traffic separately from ChatGPT traffic in GA4?

Perplexity traffic appears in GA4 as perplexity.ai / referral, a consistent source value that is easy to isolate. Create a separate channel entry in your custom channel group for Perplexity with condition: Session source exactly matches perplexity.ai. This lets you compare ChatGPT vs Perplexity traffic volume, engagement rate, and conversion rate independently. For most content sites, Perplexity referral traffic has higher engagement than ChatGPT referrals because Perplexity users click through to verify sources more consistently.

GA4 shows organic, direct, referral, and paid channels by default. It does not show a ChatGPT channel or an AI search channel. But that traffic is already in your data; it is just sitting under referral with source values that most analytics setups do not filter for. Analytics Mania’s research on AI traffic in GA4 confirms that most teams are missing a significant portion of AI-referred sessions due to incomplete source matching. How to track chatgpt traffic in analytics means knowing the exact source/medium values each AI platform passes, building a custom channel group that surfaces all of them together, and understanding that a portion of AI-referred traffic will never appear as referral because the referrer header is stripped before it reaches your server. This post is part of the full guide on AI for SEO automation.

How to Track ChatGPT Traffic in Analytics: The Platform Referrer Map

Direct Answer: How to track chatgpt traffic in analytics requires building a GA4 custom channel group using a regex that matches source values for all major AI platforms: chatgpt.com and chat.openai.com for ChatGPT, perplexity.ai, claude.ai, gemini.google.com, and edgepilot for Bing Copilot. A portion of AI traffic will always appear as direct due to referrer suppression from copy-paste navigation.

The per-platform referrer map (the reference table no other guide publishes):

AI PLATFORM          SOURCE IN GA4                  MEDIUM    NOTES
ChatGPT (current)    chatgpt.com                    referral  Post-2024 domain migration
ChatGPT (legacy)     chat.openai.com                referral  Pre-2024 sessions, still active
Perplexity           perplexity.ai                  referral  Consistent, easy to isolate
Claude.ai            claude.ai                      referral  Anthropic consumer product
Gemini               gemini.google.com              referral  Google AI chat
Bing Copilot         bing.com                       referral  Often blends with Bing organic
Bing Copilot (Edge)  edgepilot.localhost            referral  Edge sidebar Copilot sessions
Grok                 x.ai                           referral  Rare but trackable

This table is the starting point for building your custom channel group. Most guides provide only ChatGPT and skip Bing Copilot’s Edge-specific source value (edgepilot.localhost), which means a significant portion of Copilot-referred sessions remain invisible. For how AI traffic volume connects to the broader automation reporting workflow, see how to build an AI SEO dashboard.

Step 1: Build the GA4 Custom Channel Group

How to track chatgpt traffic in analytics starts with creating a channel group that surfaces all AI platforms as a single trackable source.

Step-by-step in GA4:

Go to Admin (gear icon, bottom left)
Under Data Display, click Channel Groups
Click New Custom Channel Group
Name it: “AI Search Traffic”
Click Add new channel
Name the channel: “AI Platforms (All)”
Set condition: Session source matches regex:

chatgpt\.com|chat\.openai\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|edgepilot|x\.ai

Save the channel, then save the channel group
The new channel appears in Reports → Acquisition → Traffic Acquisition within 24-48 hours. For how to schedule automated monthly AI traffic reports via n8n, see how to set up SEO automation with n8n.

For per-platform breakdown, add separate channels before the “all” group:

Channel: "ChatGPT"
Condition: Session source matches regex  chatgpt\.com|chat\.openai\.com

Channel: "Perplexity"
Condition: Session source exactly matches  perplexity.ai

Channel: "Claude AI"
Condition: Session source exactly matches  claude.ai

Channel: "Bing Copilot"
Condition: Session source matches regex  edgepilot

Channel: "AI Platforms (All)" [catch-all]
Condition: Session source matches regex [full regex above]

With this structure, each platform appears as its own row in the Traffic Acquisition report, plus the catch-all row shows total AI traffic. Changing the channel group dropdown in the report from “Default Channel Group” to your custom group shows the breakdown immediately.

Step 2: Build the GA4 Explore Segment

The channel group shows AI traffic in the Acquisition report. The Explore segment lets you analyze AI traffic behavior: conversion rate, funnel completion, cohort comparison against organic users. This is the step every other guide misses.

Creating the AI traffic segment in Explore:

Go to Explore (compass icon, left nav)
Create a Free Form exploration
Click the + next to Segments
Choose Create a new segment
Select Session segment (not User segment)
Add condition: Session source matches regex chatgpt\.com|chat\.openai\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|edgepilot|x\.ai
Name it: “AI Platform Sessions”
Save and apply to your exploration

Now add a second segment for comparison: Organic Search (use the default system segment). This gives you a side-by-side view of AI-referred vs organic session behavior across engagement rate, conversion rate, pages per session, and session duration. For most content sites, AI traffic converts at a lower rate but engages at a higher depth per session, which confirms AI users arrive with specific intent rather than broad browsing behavior.

Step 3: Identify and Quantify Dark Traffic

Not all AI traffic appears as referral. How to track chatgpt traffic in analytics fully means accounting for the traffic that does not pass a referrer header.

What dark traffic is:

When a user reads a ChatGPT or Perplexity response that cites your URL, then copies the URL and pastes it into a new browser tab, no referrer header is sent. GA4 records this session as direct traffic with no source. This is referrer header suppression and it is the primary reason AI-referred traffic is systematically undercounted in every analytics account.

The dark traffic proxy signals:

Dark traffic from AI citations produces distinctive patterns in the direct channel. In GA4, navigate to Reports → Acquisition → Traffic Acquisition, filter to Session default channel group = Direct, then look for:

DARK TRAFFIC INDICATORS IN DIRECT CHANNEL:
New user ratio above 40%     — Direct traffic is usually returning users who know your URL
                               A high new user spike in direct suggests copy-paste arrivals
Session duration under 10s   — Genuine branded direct visitors browse; paste-arrivals often
                               arrive on a specific page and leave without exploring
Landing page = content page  — Direct traffic usually lands on the homepage or known pages
                               AI-cited direct traffic lands on specific content articles
Spike coincides with          — If you confirmed a citation in ChatGPT on Tuesday and direct
AI citation event              traffic to that URL spiked on Wednesday, that is dark traffic

The UTM workaround for internal linkings:

If you know a specific URL is being cited in AI Overviewss or Perplexity (check via how to track AI Overview impressions in GSC), add UTM parameters to the internal links on that page: ?utm_source=ai_citation&utm_medium=referral&utm_campaign=aio. When a user arrives via the AI-cited URL and clicks to another page on your site, the UTM parameter carries forward in the session, and the converted traffic is attributed to AI citation rather than direct or organic.

The AI Traffic Insight Most Teams Miss

Most teams set up AI traffic tracking to measure volume: how many sessions per month from ChatGPT, what is the trend, is it growing. How to track chatgpt traffic in analytics at a strategic level means asking a different question: which of your pages are getting AI traffic and what made those pages citable?

Pull the AI Platform Sessions segment in GA4 Explore. Add the Landing Page dimension. The pages getting AI-referred traffic are the pages your content formats (direct answer blocks, FAQ sections, clear entity headings) are working on. Those are your citation-eligible pages. Build more content on the same topics, using the same formats. The AI traffic report is a content strategy signal, not just an analytics metric.

“Your AI traffic report shows you which content the AI systems have already decided is worth citing. That is your highest-confidence content expansion signal.”

The second insight: compare engagement rate between your AI Platform Sessions segment and your Organic Search sessions segment. If AI traffic has a higher engagement rate (users reading more pages per session), it means your content is satisfying AI-referred users who arrive with specific intent. If AI traffic has a lower engagement rate, users are arriving from an AI citation but not finding what they expected; that is a content relevance gap the cited page does not fully cover.

“If AI traffic bounces harder than organic, the cited page answers the AI’s question but not the user’s follow-up question. Add that follow-up as the next section.”

Where AI Traffic Tracking Breaks

Failure 1: Using only chatgpt.com in the channel group and missing chat.openai.com. ChatGPT migrated to chatgpt.com as the primary domain but the legacy chat.openai.com domain still generates referrals. A channel group that only matches chatgpt.com misses all legacy-domain sessions. Use the combined regex chatgpt\.com|chat\.openai\.com for the ChatGPT channel. Run a report filtered to Session source contains “openai” to see how many sessions you currently have from the legacy domain.

Failure 2: Treating all dark traffic as unresolvable. Dark traffic can be partially recovered through the UTM-in-internal-links approach and through direct traffic anomaly analysis. Teams that dismiss dark traffic as unknowable miss the opportunity to attribute a meaningful portion of direct traffic spikes to AI citations. The proxy signals (new user ratio, session duration, landing page pattern, timing correlation with confirmed citations) together identify dark traffic with reasonable confidence even without a referrer header.

Failure 3: Not comparing AI traffic engagement to organic traffic engagement. Setting up the channel group and then only looking at session volume misses the behavioral insight. AI-referred traffic has a different conversion pattern than organic search traffic because users arrive from a specific cited passage, not from a query intent. Without the Explore segment comparison, you have a volume number but no understanding of what AI traffic means for your business outcomes.

Failure 4: Having AI crawlers blocked in robots.txt and then diagnosing low AI traffic. A site that has User-agent: GPTBot / Disallow: / in robots.txt will see almost no chatgpt.com referral traffic regardless of content quality. The JavaScript rendering connection matters here too: if your site is client-side rendered and AI crawlers cannot read your content, see what is javascript rendering for seo and ai before diagnosing AI traffic volume. Check yoursite.com/robots.txt before building AI traffic tracking infrastructure. If AI crawlers are blocked, the traffic tracking will show near-zero volume and every diagnostic will be misleading. Fix the robots.txt disallow first; then set up tracking to measure the baseline.

Frequently Asked Questions

Four questions on how to track chatgpt traffic in analytics answered directly:

What is the source/medium for ChatGPT traffic in GA4?
Is AI traffic counted as direct traffic in GA4 by default?
How do I create a custom channel group for AI traffic in GA4?
How do I track Perplexity traffic separately from ChatGPT traffic in GA4?

What is the source/medium for ChatGPT traffic in GA4?

ChatGPT traffic appears as chatgpt.com / referral for current sessions and chat.openai.com / referral for sessions from before the 2024 domain migration. How to track chatgpt traffic in analytics accurately means your custom channel group regex must include both values. Using only chatgpt.com will miss a significant number of sessions still arriving via the legacy domain. Check your GA4 Acquisition report filtered to Session source contains “openai” to quantify the volume you are currently missing.

Is AI traffic counted as direct traffic in GA4 by default?

Some AI traffic does appear as direct. Users who copy and paste a URL from a ChatGPT or Perplexity response into a new browser tab trigger no referrer header, so GA4 records the session as direct. The proxy signals for this dark traffic are: high new user ratio in the direct channel above 40%, very short session duration under 10 seconds, landing pages matching AI-cited content pages, and timing correlation with a confirmed AI citation. None of these signals alone is conclusive, but together they identify dark traffic sessions with reasonable confidence.

How do I create a custom channel group for AI traffic in GA4?

Go to Admin in GA4, then Data Display, then Channel Groups, and create a new custom channel group. Add a channel called AI Platforms with Session source matching the regex chatgpt\.com|chat\.openai\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|edgepilot|x\.ai. Save and wait 24-48 hours. The channel appears in the Traffic Acquisition report. For per-platform breakdown, add separate channels for ChatGPT, Perplexity, and Claude before the catch-all channel, using exact source matches for each.

How do I track Perplexity traffic separately from ChatGPT traffic in GA4?

Add Perplexity as its own channel in the custom channel group: condition Session source exactly matches perplexity.ai. This isolates Perplexity sessions from the aggregated AI channel. How to track chatgpt traffic in analytics with per-platform clarity means placing the specific platform channels above the catch-all channel in the channel group order; GA4 assigns sessions to the first matching channel rule, so more specific rules must precede broader rules. With per-platform channels in place, your Traffic Acquisition report shows separate rows for ChatGPT, Perplexity, Claude, and the combined catch-all.

Before setting up AI traffic tracking, run these five checks:

Does your robots.txt block GPTBot, PerplexityBot, or ClaudeBot? (Check yoursite.com/robots.txt; if any AI crawler is blocked, you will have near-zero AI referral traffic regardless of tracking setup)
Do you have both chatgpt.com and chat.openai.com in your channel group regex? (Missing the legacy domain means undercounting ChatGPT referrals from pre-2024 sessions)
Have you created a GA4 Explore segment for AI Platform Sessions? (The Traffic Acquisition report shows volume; the Explore segment shows behavior and conversion comparison with organic)
Have you identified which specific pages receive the most AI-referred traffic? (Landing page dimension in the Explore segment reveals your citation-eligible content, your content expansion map)
Have you noted the baseline AI traffic share for this month to compare against next month? (AI traffic as a percentage of total organic + referral sessions is the trend metric that matters, not absolute session count)

That is how to track chatgpt traffic in analytics as an operational system rather than a one-time configuration. If you want help setting up the full AI traffic tracking and attribution workflow including dark traffic identification and content strategy signal analysis, my AI SEO services include the complete GA4 AI attribution setup.