The management of one-to-one business interaction is challenged by the latency in the acquisition of information about the individual customer's preferences. Although sharing this type of information would empower service providers to personalize the interaction with new customers since the first connection, this idea can be hardly applied in real cases if the service provider cannot protect the data it has acquired from competitors and select the trusted parties from which it wants to receive information. As a solution, we propose a framework supporting the controlled sharing of customer information between e-commerce applications. Our framework includes two main components: 1) a Trust Management System (running off-line with respect to the information sharing service), which enables the service provider administrator to specify restrictions on the service providers to be considered as trusted parties; 2) a User Modeling Agent, which manages the propagation of customer data betwe...