A Model-Based Customer Inference Engine

In complex and frequently changing product domains, customers need qualified consultation services that allow them to make well-informed purchasing decisions without requiring excessive technical knowledge about the product domain. In this paper we present an intuitive customer metamodel with an associated inference engine which can be integrated into conversational recommender systems to provide comfortable reasoning about customer needs. The inferred knowledge can be used to manage the recommendation dialogue and to map non-technical customer statements into preferences for technical attributes.