This paper reports on our development of a conceptual model, a design exemplar, of a purposeful, inhabitable, intelligent architectural setting, spatially reconfigured by means of robotics. Our novel conceptual model, CoPRA, is inspired by Christopher Alexander’s notion of a “Compressed-Pattern Architecture,” in which a single space is reorganized to become many different, functional rooms. In our exemplar, this reorganization is not performed by inhabitants, manually, but instead by robotics actuated in response to human activity as sensed by an embedded computational system. Whether the spatial reorganization of CoPRA occurs intelligently or interactively, it occurs through precise, purposeful control in support of human needs and opportunities. CoPRA was developed through a careful process defined by “Research Through Design” [RTD], an emerging methodology in human-computer interaction and design research that promises to prove productive to designing intelligent architectural settings. More broadly, RTD and Compressed-Pattern Architecture represent two productive conceptual tools for a growing research community focused at the interface of architecture and computing. Our own interdisciplinary team is currently applying these two means, along with the CoPRA exemplar, in our design, prototyping and evaluation of a functional, cyber-physical library module, room-scaled, delivered to the interior of an existing branch public library serving an underserved community that would otherwise not receive library facilities exhibiting much in the way of functional, technological or design sophistication.