Conceptual Design is generally performed using historical data and low-order numerical models due to their computational expedience and low threshold of input information. However, performance predictions provided by these tools sometimes fail to capture critical effects due to a lack of interdisciplinary coupling, insufficient fidelity within a single fidelity, or the omission of a discipline altogether. In the best case, this inadequacy leads to missed design opportunities, but also may result in costly, late-stage design corrections or the production of a vehicle with limited capability in the worst case. While recent efforts have made progress toward bringing higher-order, physics-based predictions forward in the design process, use of these tools has been hindered by the modeling tools typically employed. The Computational Aircraft Prototype Syntheses research program attempts to provide a modeling infrastructure that enables agile, physics-based design by analysis, decoupling the availability of certain analysis tools from the stage of design. This manuscript describes a shift in thinking about vehicle modeling and geometry definition and its integration with meshing and analysis, and provides examples of applications in the literature. Ultimately, multiple, consistent, analysis-specific geometries should be outputs of a unifying design model, and should play an active role throughout the entire analysis process.