AI Agent-Assisted Design Method for Partial Interior Renovation of Existing Homes

Addressing the challenges associated with partial renovation of existing residential interiors, including ordinary homeowners’ incomplete expression of design needs, limited access to professional design support, and the inability of conventional AIGC tools to accurately interpret complex design requirements, this study proposes an AI Agent-assisted design method for partial interior renovation. An AI Agent workflow was developed to establish an integrated process encompassing requirement acquisition, design semantic mapping, and design-scheme generation. The workflow automatically transforms natural-language requirements into structured design information and improves the quality and consistency of generated designs through multimodal information analysis and prompt-weight optimization. Experimental results show that the AI Agent-assisted method achieves improvements in requirement alignment, spatial structure preservation, spatial aesthetics, and generation stability. SUS analysis indicates that this method can provide relatively accessible design decision support for non-professional users. Overall, this study demonstrates that AI Agent-assisted human–AI collaborative design has the potential to support homeowners in independently designing and refining partial residential interior renovation schemes. Although this study did not directly measure indicators such as material consumption, embodied carbon, or service-life extension, the proposed workflow may also offer a potential decision-support pathway for more efficient and incremental renovation of existing homes.