Full Answer

The core problem is that AI-driven traffic looks invisible in standard analytics. When ChatGPT or Perplexity cites your content, the resulting visit often arrives with no referrer or gets bucketed as Direct in GA4 — making AI traffic indistinguishable from someone typing your URL directly. This isn't a bug in your analytics; it's a structural gap in how AI platforms handle outbound links.

Three data sources, combined, close most of that gap. First, Google Search Console's AI Mode filter shows which of your pages appear in AI-generated answers and how many impressions and clicks they receive — the only free tool that measures Google's own AI citations directly. Second, server log analysis identifies AI crawler visits by user-agent string: GPTBot for OpenAI, Google-Extended for Gemini, and ClaudeBot for Anthropic. These crawl patterns reveal which pages AI engines are actively indexing for citation.

Third, server-side tracking recovers referrer data that client-side JavaScript drops. When an AI platform sends a visitor, the HTTP referrer header often arrives intact at the server level even when the browser strips it before GA4's JavaScript fires.

Here's the thing... the businesses getting this right aren't waiting for a single tool to solve it. They're triangulating across all three sources — and the ones that start now build the baseline data needed to measure AI channel growth over time.