—Reasoning by interrogation is one of the most ancient and experimented ways of reasoning. Originated by the Aristotelian elenchus, it has been used for many purposes, such as the resolution of mathematical and daily problems [25], [26], the discovery of new knowledge [19], [34], [36], the realization of questioning/answering processes [23]. In this paper we present the conceptual foundations of interrogative agents, a new model of BDI architecture based on interrogative logic. This model allows us to express the properties of agents in a natural way, and to use heuristics for reasoning. Finally, in order to explicate the whole approach and to highlight its main features we describe the application of interrogative agents in the context of database refactoring.