Across many generations, including the digitally naïve and savvy, digital technologies introduce something other to the design process. This other quality ranges from inaccessibility to facile but opaque production of complexity, yet in all cases has obscured the recognition of authorial control. Fundamentally, my research interrogates this loss of control, or confusion of agency. Using the architectural model as a site of inquiry, I seek out sources of authorial control in digital design and ultimately speculate productive forms of digital agency for architecture. Digital models should be understood as continuing the legacy of architectural models rather than as a novel product of technology. As a mediator of communication, all models represent in increasing levels of abstraction. Understood literally, models mediate as images. In the formalization of underlying structure, models mediate as diagrams. Formulating structure as abstract logic, models mediate as mathematics. Digital models continue to mediate in these manners, but the nature of digital media places a new emphasis on the precise, computable definition of all of these roles. The digital process allows an author to construct a creative dialogue with a model. In this way, models are tools which operate on patterns of thought by creating feedback loops in the model- author dialogue. Formalized as the return of output back to a process as input, feedback is a mechanism which may structure both architectural form and the design decision-making process itself. In complex natural systems, feedback explains the formation of organizational patterns. In cybernetics, feedback explains the auto-regulation of an organized system. All forms of feedback can be understood as the cyclical communication of information. As such, processes of transmission and translation play a crucial role in the conveyance of information from source to receiver and back. Applied to digital model-making, feedback becomes an active decision-making mechanism, allowing an author to construct an apparatus to aid evaluation of work-in-progress and definition of increasingly specific intents. Appropriating the techno-industrial concept of design intent and citing the computational mathematical basis of digital models, I speculate the existence of a non-metaphorical solution space which can be instrumentalized by digital designers. The formalized yet iteratively changing set of possible solutions to an interpreted design problem is termed “intention space.” Computational codification of design intent makes mathematical descriptions of a digital model accessible and affords the possibility for a directed search thorough the model’s highly-structured intention space. Intention space demystifies the digital design process enabling architects to directly manipulate digital media both intuitively and deliberately. The idea of intention space is explicitly contrasted with two prevailing theories of control in digital methods: authorless design and split agency. Unlike these assertions, the encoding of design intent through mechanisms of constructed feedback and visualized as intention space proffers digital technologies as uniquely enabling instruments within the creative process.