Over the course of our studies of how intelligence analysis professionals perform their work, we have observed that the perceived utility of technosocial models varies with the degree that analysts could create and/or adapt a model to a particular problem or situation. Analysts failed to trust models (a form of validity) that were created outside of their community of practice. In this paper, we present relevant results from our studies as they pertain to how analysts might create or adapt technosocial models. In particular, we define a number of significant concerns and challenges that must be overcome to enable any sort of user-created or user-adaptable modeling approach. We present a set of tools and techniques that have emerged from our own experiences developing model-based analysis and decision-support systems to illustrate the core challenges and suggest some potential solutions.