Amazon knows what customers do  every click, search, and purchase  but that data is fragmented across experiences. No single place tells you what a customer is trying to accomplish in the moment. A runner training for his first marathon browses running shoes on the Amazon app, then asks Alexa whether marathon-specific shoes hold up over high mileage. These are two disconnected signals in two different systems  nothing connects them into one story: this customer is training for a marathon and needs gear built to go the distance. As a result, every AI experience has to reconstruct intent from raw events on its own — slow, expensive, and inconsistent and the customer's intent ends up understood shallowly, late, and differently on every surface.