Exploring the Convergence of AI and VR in Architectural Design, From Ideation to Actionable Outputs

The integration of Artificial Intelligence (AI) and Virtual Reality (VR) presents new opportunities for transforming architectural design workflows, particularly in small-scale projects. This research introduces MerGen, a prototype system that combines generative AI models, large language models (LLMs), and immersive VR to support real-time collaboration between designers and intelligent agents. By situating users within panoramic environments, MerGen enables continuous spatial awareness and interactive control throughout the design process.

Generative models like Stable Diffusion allow for real-time visual modifications based on user intent, while LLMs provide contextual suggestions, semantic dialogue, and design documentation. This workflow bridges conceptual ideation with actionable outputs, including material recommendations and spec-style summaries.

A modular, experiment-driven methodology is used to evaluate MerGen’s effectiveness across two key stages of design: early sketching and thematic exploration, and material selection and specification. The study examines how AI–VR integration enhances ideation, visualization, and decision-making. Findings highlight both the promise and limitations of this approach, offering insights into its role in supporting immersive, multimodal architectural workflows.