Zero-Click Attribution: Measuring Dark AI Search Pipeline in B2B RevOps

How to track, attribute, and prove ROI from buyers who research your brand inside ChatGPT and Claude before typing your URL directly into their browser.

[ EXECUTIVE RESEARCH BRIEF // CITADEL SEARCH LAB ]

Combining HTTP referrer rules (chatgpt.com / perplexity.ai), server-side bot log telemetry, and self-reported intake fields reveals 38% hidden AI pipeline.

[ PART 01 // EMPIRICAL COHORT TELEMETRY & CITATION DATA ]

Standard Google Analytics 4 setups misclassify up to 60% of AI-originated buyer journeys as 'Direct' traffic because enterprise buyers read a synthesis inside ChatGPT Desktop or Claude, open a new tab, and navigate straight to your domain.

[ PART 02 // PRODUCTION ARCHITECTURE & JSON-LD BLUEPRINT ]

The Citadel Attribution Protocol combines three layers: (1) Edge server log monitoring of OAI-SearchBot, PerplexityBot, and ClaudeBot crawl frequency on high-intent pages; (2) UTM and Referrer parsing for `chatgpt.com` and `perplexity.ai`; and (3) A calibrated 'How did you evaluate us?' intake field paired with 150-prompt Share-of-Voice tracking.