How to Track Traffic From ChatGPT and AI Search in Google Analytics 4

Why AI traffic is hard to see

When someone clicks a cited source in ChatGPT, Perplexity, Gemini or Copilot, they land on your site like any other visitor. GA4 classifies most of those visits by their referrer — and because GA4’s default channel rules don’t include an AI category, they end up mixed into Referral, alongside directories and partner links.

That makes it almost impossible to answer the question every business owner is now asking: is AI search actually sending us customers? The fix takes about fifteen minutes.

The AI referrers to look for

Step by step: create an “AI Assistants” channel

1. Open Admin → Data display → Channel groups

   In GA4, go to Admin and, under Data display, choose Channel groups. Copy the default channel group to create a new custom group.

2. Add a new channel called “AI Assistants”

   Set the condition to Source matches regex, and paste the expression below.

3. Move it above “Referral”

   Channel rules are evaluated in order. If AI Assistants sits below Referral, the visits will still be counted as Referral.

4. Save and use it in reports

   In Traffic acquisition, switch the primary dimension to your custom channel group. It also applies retroactively to historical data.

(chatgpt\.com|chat\.openai\.com|perplexity\.ai|gemini\.google\.com|copilot\.microsoft\.com|claude\.ai)

Once the channel is in place, give it a few weeks before drawing conclusions. Many sites see only a handful of AI visits a week at first. What matters is the direction of travel — and whether those visitors go on to enquire.

Measure value, not just visits

Sessions alone don’t tell you much. Once the channel exists, compare it with your other channels on the metrics that matter:

Build a simple monthly AI report

A one-page monthly report is enough for most businesses. Keep the same format each month so trends are easy to see:

1. AI sessions and key events

   From your AI Assistants channel — compared with last month and the same month last year.

2. Top AI landing pages

   Which pages AI engines send people to. These are your most quotable assets — protect and expand them.

3. Brand search trend

   From Search Console. Rising brand searches often reflect AI mentions that didn’t produce a direct click.

4. Citation snapshot

   How often you’re named for your fixed prompt set, by engine.

To see landing pages for AI traffic, open Reports → Engagement → Landing page, add a comparison for your AI Assistants channel, and sort by sessions. For deeper analysis, a GA4 Exploration with Session source and Landing page as dimensions shows exactly which pages each assistant sends visitors to.

Turning AI traffic data into action

Numbers only matter if they change what you do. Three patterns are worth acting on. If one page attracts most AI visits, it’s working as a citable resource — expand it, keep it current and link your service pages from it. If AI visitors land on informational pages but rarely enquire, add a clear next step on those pages. And if AI traffic is flat while competitors are being cited, the issue is usually visibility rather than tracking — which is where crawler access, structured data and answer-first content come in.

Common mistakes when tracking AI traffic

• Placing the AI channel below Referral — the visits never reach it.
• Matching too broadly — a regex like .*ai.* catches unrelated sites containing “ai”.
• Forgetting old domains — chat.openai.com still appears in historical data.
• Reporting sessions without conversions — the point is enquiries, not visits.
• Judging one month in isolation — volumes are small and noisy; look at three-month trends.

What GA4 can’t show you

• Visits with no referrer — some apps and copy-pasted links arrive as Direct.
• Influence without clicks — a customer who reads an AI answer, then searches your brand on Google, shows up as organic or direct.
• AI Overviews clicks — blended into Google organic traffic.
• Whether you were mentioned at all — that needs citation tracking with a fixed prompt set.

That’s why we pair analytics with monthly prompt-based citation tracking. Analytics shows the clicks; citation tracking shows the visibility that drives them. Our GEO methodology explains how we measure both.

About the author

Dr Andrej Miklosik

Founder of Catnapweb and Professor at the Department of Marketing and Commerce, Comenius University. Author of 170+ research publications on AI in marketing and search visibility, cited 1,705 times. Full profile →