Most teams don’t know and it’s understandable.
LLMs share almost no data, and the companies trying to “model” LLM behavior can’t prove how accurate they are.
So how can you tell what’s really happening? LLMs send two signals your site can detect:
1️⃣ AI chatbot searches: when a chatbot performs a web search, your site logs it.
2️⃣ LLM referrals: when an LLM recommends your site and a user clicks through.
GA4 can give you referral data but GA4 uses JavaScript tracking, and AI chatbots don’t run JavaScript. That means a huge amount of AI-driven activity is invisible to Google Analytics, Mixpanel, Amplitude, and every other JS-based analytics platform.
To know how you’re actually performing in LLMs, you need to:
- Log traffic at the edge
- Parse raw logs for AI bot and referral traffic
This is the only dependable method for measuring how your brand is represented inside AI chatbots.