[How To] Google Analytics Tracking of AI Engine Referrals to Your Website

You can see traffic from LLMs (ChatGPT, Perplexity, Claude, Gemini, etc.) in Google Analytics 4 (GA4), but it’s a bit hidden because most LLM referrals don’t show up as a clean “LLM” channel. Here’s how to surface it clearly.

1. Where LLM traffic usually appears in GA4

LLM traffic can show up as:

• Referral traffic (most common)

• Organic Search (especially Google Gemini / SGE)

• Direct (when referrer is stripped)

So you’ll need to look beyond default channel groupings.

2. Quick win: View LLM referrals directly

Step-by-step

1. Go to Reports → Acquisition → Traffic acquisition

2. Change the primary dimension to Session source / medium

3. Use the search box or filter for known LLM sources:

Common ones to look for:

chat.openai.com
openai.com
perplexity.ai
claude.ai
anthropic.com
bard.google.com
gemini.google.com
copilot.microsoft.com
bing.com (Copilot traffic often hides here)
you.com
phind.com

Most will appear as:

chat.openai.com / referral
perplexity.ai / referral

3. Create an Exploration for deeper analysis (recommended)

This is the cleanest way to track LLM traffic over time.

Steps

1. Go to Explore → Blank

2. Add dimensions:

   • Session source

   • Session medium

   • Landing page

3. Add metrics:

   • Sessions

   • Engaged sessions

   • Conversions (if set)

4. Add a Filter:

   • Include → Session source → contains →openai | chat | perplexity | claude | anthropic | gemini | copilot | you.com | phind

Now you’ll see:

• Which pages LLMs send traffic to

• Engagement quality vs search/social

• Conversion performance

4. Identify “hidden” LLM traffic inside Organic Search

Some AI answers (especially Google SGE / Gemini) still count as organic search.

To isolate this:

1. Go to Search Console → Performance

2. Look for:

   • Very long, conversational queries

   • Question-based searches that align with AI-style prompts

3. Compare landing pages that spike without traditional SEO changes

This isn’t perfect, but it helps identify AI-assisted discovery.

5. Optional: Create a custom “LLM” channel group

If you want LLM traffic visible in standard reports:

1. Go to Admin → Data settings → Channel groups

2. Create a new channel called LLM

3. Rules:

   • Session source contains:chat.openai.com perplexity.ai claude.ai anthropic.com gemini.google.com copilot.microsoft.com you.com phind.com

This lets LLM traffic show up alongside Organic, Paid, Social, etc.

6. Why LLM traffic is often underreported

• Some LLMs strip referrer data

• In-app browsers often show as Direct

• Copilot traffic frequently appears as bing.com

• AI summaries may not generate a click at all (zero-click exposure)

7. Bonus: Best practices to improve LLM-attributed traffic

To get more measurable traffic from LLMs:

• Add clear citations and authoritative language

• Use FAQ-style content

• Publish original data or definitions

• Make sure pages are crawlable and fast

• Add schema (FAQ, HowTo, Article)

TL;DR

• Look under Referral → Session source

• Build an Exploration with LLM source filters

• Create a custom LLM channel group

• Expect some traffic to hide under Direct or Organic