The recent increase in interest in Interactive Storytelling systems, spurred on by the emergence of affordable virtual reality technology, has brought with it a need to address the way in which narrative content is visualized through the complex staging of multiple narrative agents’ behaviors within virtual story worlds. In this work we address the challenge of automating several aspects of staging the activities of a population of narrative agents and their interactions, where agents can have differing levels of narrative relevance within the situated narrative actions. Our solution defines an approach that integrates the use of multiple dynamic regions within a virtual story world, specified via a semantic representation that is able to support the staging of narrative actions through the behaviors of the primary and background agents’ that are involved. This encompasses both the mechanics of dealing with the narrative discourse level as well as the interaction with the narrative generation layer to account for any dynamic modifications of the virtual story world. We refer to this approach as mise-en-scène region (miser) support. In this paper, we describe our approach and its integration as part of a fully implemented Interactive Storytelling system. We illustrate the work through detailed examples of short narrative instantiations. We present the results of our evaluation which clearly demonstrate the potential of the miser approach, as well as its scalability.