How to Track AI Traffic in GA4 (ChatGPT, Perplexity, Gemini & Copilot) — 2026 Setup

AI is answering your customers' questions before they ever open a tab.

When AI uses web sources, it often shows citations/links users can click:

• ChatGPT search responses include inline citations and a Sources view. (OpenAI Help Center)
• Perplexity says each answer includes numbered citations linking to original sources. (Perplexity AI)

So yes — AI can send traffic. But most teams don't see it because GA4 lumps those visits into Referral and hides them in the noise.

This guide gives you a clean, repeatable setup to:

• isolate AI referrers,
• put them in a dedicated "AI Chatbots" channel,
• and see which pages AI actually sends people to.

No fluff. You can do it in ~20 minutes.

---

What you'll build (so your reporting stops lying)

You'll create two things:

1. An Exploration report that lists AI sources + landing pages
2. A custom channel group in GA4 so "AI Chatbots" shows up next to Organic / Paid / Referral

Google notes that GA4's default channel group can't be edited, but you can create custom channel groups with your own rules. (Google Analytics Help)

---

Step 1 — Quick check: confirm AI referrals exist

In GA4 go to:

Reports → Acquisition → Traffic acquisition

Add a filter:

• Session source / medium contains chatgpt Then repeat for:
• perplexity
• gemini
• copilot

If you see anything, great — you're just not labeling it properly yet.

If you see nothing, that's also useful: you're early (or your niche isn't being cited yet).

---

Step 2 — Build an "AI Traffic" Exploration (best for analysis)

Explorations are where you'll answer the real questions:

• Which AI source sends the most engaged visits?
• Which landing pages get AI traffic?
• What content types pull AI clicks?

A very practical way to do this is outlined step-by-step by Loves Data (including reusable Exploration setups). (Loves Data)

Recommended Exploration layout:

• Rows: Session source → Landing page + query string
• Values:
  • Sessions
  • Engaged sessions
  • Engagement rate
  • Conversions (if configured)
• Filter: Session source matches regex (next step)

---

Step 3 — Add a regex filter to catch the main AI referrers

In your Exploration, add a filter on Session source (or Session source / medium) and use a regex.

A solid starting regex pattern (commonly used in GA4 channel grouping examples) is documented by Loves Data, including ChatGPT + Gemini + Perplexity + Copilot. (Loves Data)

Starter regex (safe + readable):

chatgpt|openai|perplexity|gemini|bard|copilot

Then expand once you verify what actually shows up in your data.

Pro tip: keep a "Known AI Referrers" note and update it quarterly. Referrer domains change.

---

Step 4 — Create an "AI Chatbots" channel group (for executive reporting)

This is the part that makes GA4 dashboards usable.

Go to: Admin → Data display → Channel groups

Google's help docs explain channel groups and confirm you can create custom ones (since defaults aren't editable). (Google Analytics Help)

Create:

• Custom channel group name: AI Chatbots
• Add a channel inside it: AI Chatbots
• Rule example:
  • Field: Session source
  • Condition: matches regex
  • Value: your AI regex (from Step 3)

Now "AI Chatbots" becomes a first-class channel you can trend month-over-month.

(If you want more tactical GA4 UI walkthroughs for channel grouping mechanics, guides like Online Metrics cover the admin flow and rule structure.) (Online Metrics)

---

Step 5 — Find the pages AI is sending traffic to (the real gold)

Once your Exploration is working, sort your rows by:

• Sessions (descending) Then look for patterns:

You'll usually see AI traffic cluster around:

• "definition" pages ("what is…?")
• comparisons ("X vs Y")
• alternatives ("alternatives to…")
• checklists ("how to choose…")
• stats/benchmarks (if you have original data)

Those are your AI entry pages. Upgrade them first:

• add an answer block under each H2
• add a small table
• add citations to primary sources

(That's how you turn "some traffic" into "compounding traffic.")

---

Step 6 — Stop misattribution (the 4 common GA4 mistakes)

1) Confusing source vs medium

AI visits are usually referral medium, but you should group by source domain (that's the stable identifier).

2) Losing referrers due to redirects

If you bounce users through tracking links or intermediate domains, you can accidentally break referral tracking.

3) Looking only at last-click conversions

AI clicks often assist conversions (people research, then come back later). Track assisted conversions where you can.

4) Treating "AI traffic" as one bucket forever

Different AI sources behave differently. Keep the channel group for reporting, but analyze sources separately in Explorations.

---

What GA4 still can't tell you (and why AI visibility ≠ AI traffic)

GA4 can tell you:

• who clicked
• from where
• to what page
• what they did next

GA4 cannot tell you:

• how often AI mentions you without a click
• whether AI recommends competitors instead
• whether AI describes you accurately

That's the gap between traffic measurement and AI visibility.

---

Next step

If you implement this, you'll finally know:

• which AI source sends you users,
• which pages earn those clicks,
• and what content formats work.

If you want to go one level deeper (the level that actually wins markets): you measure recommendations + citations + positioning accuracy across the prompts your buyers use.

---

Sources

• Google Analytics Help — Default channel group / custom channel groups in GA4: https://support.google.com/analytics/answer/9756891?hl=en
• Loves Data — How to Track AI Traffic in GA4: https://www.lovesdata.com/blog/how-to-track-ai-traffic-ga4/
• OpenAI Help Center — ChatGPT search: https://help.openai.com/en/articles/9237897-chatgpt-search
• Perplexity Help Center — How does Perplexity work?: https://www.perplexity.ai/help-center/en/articles/10352895-how-does-perplexity-work
• Online Metrics — Custom channel grouping in GA4: https://online-metrics.com/custom-channel-grouping-in-ga4/