TL;DR
Build AI attribution in 4 layers: crawler intelligence (are AI bots reading your pages?), citation monitoring (are you being cited?), traffic attribution (are citations driving visits?), and revenue attribution (are visits converting to sales at the SKU level?).
The invisible revenue channel
A customer asks ChatGPT for a product recommendation. Your brand is cited. They click through and buy. In your analytics, this shows up as "Direct" or "Referral." No keyword. No campaign. No attribution trail.
AI search is driving purchase decisions and the revenue is landing in the unattributed bucket of every analytics platform.
Why traditional analytics can't see AI search
Google Analytics was built for clicks, sessions, and UTM parameters. AI search breaks every assumption: no query data, no impression data, no position data, and session attribution collapse when users research in AI then return via direct navigation.
The AI attribution framework
Four layers: Layer 1 — Crawler intelligence (monitor GPTBot/ClaudeBot access in server logs). Layer 2 — Citation monitoring (track brand mentions across AI platforms). Layer 3 — Traffic attribution (segment AI referral sources in analytics). Layer 4 — Revenue attribution (connect citations to SKU-level purchases).
SKU-level attribution: the operator's edge
Channel-level attribution tells you "AI search drove X revenue." SKU-level attribution tells you which products AI recommends, which are invisible, which competitors appear instead, and whether citation changes correlate with revenue changes. This is what Nexeo was built to solve.