Substituting Information for Interaction: A Framework for Personalization in Service Encounters and Service Systems

We compare person-to-person service encounters with those in which the service provider is an information system to identify the capabilities needed to personalize a service encounter. We suggest “substituting information for interaction” as a principle that unifies these different types of encounters whenever the information needed to create value in a service system accumulates incrementally through human or automated customer interactions. We review research and practice in computer science, artificial intelligence, data mining, machine learning, and information systems design to bring an interdisciplinary robustness to our conceptual proposal. Human service providers and automated service systems both need (1) a service model manager that stores information about how a customer requests a service; (2) a customer model manager that stores information about customers and preferences; (3) a recommendation system manager that uses service models, customer models, and contextual information to adapt the service at delivery time; (4) a learning system that analyzes previous service encounters to refine service and customer models, and a (5) service monitoring system that monitors the status of service delivery. The substitution concepts and mechanisms we propose highlight the range of design choices and help managers evaluate whether a human interaction or information exchange creates or undermines value in a service system.