Desktop virtual world learning environments (DVWLEs) offer noteworthy potential for online learning, with two distinct characteristics, a sense of place and presence that differentiate them from traditional online learning platforms. These unique characteristics align DVWLEs with physical architectural design elements and principles (AEPs) and emphasize the significance of interaction design elements, particularly affordances and signifiers. Nonetheless, there is a notable research gap concerning the integration of AEPs, interaction design elements, and principles, as well as the design of affordances and signifiers, including associated design patterns, in DVWLEs’ design. This study aims to address this gap by providing conceptual frameworks for categorizing AEPs and interaction design elements in DVWLEs based on their affordances and signifiers. Additionally, we seek to identify and develop common design patterns across diverse DVWLEs, integrating them into an AI agent. This AI co-designer will empower educators to create immersive, tailored educational environments, enhancing the overall learning experience.