Architectural Experience and Artificial Intelligence: Development of A Tool to Support the Design Process

Objective: This study investigates the application of artificial intelligence (AI) in architectural experience design by developing a tool that predicts spatial perceptions based on architectural components, assisting students and architects without the need for direct user interaction. Theoretical Framework: Grounded in UX and architectural experience, the study utilizes Desmet's (2007) and Russell’s models for emotional evaluation while exploring the impact of architectural atmosphere on spatial perception. Method: Using the Design Science Research (DSR) approach, the UX Arch platform was developed as an AI-based tool. It registers architectural references and associated experiences to train a predictive model, validated by 28 students. Results and Discussion: The AI correctly predicted emotional polarity in 89% of cases and intensity in 55%, achieving an overall accuracy of 80%. Participants positively evaluated the system but suggested improvements in component categorization and database expansion. Research Implications: The study reinforces AI’s potential in predicting architectural experiences, promoting integration between technology and design while providing a systematic method for evaluating spatial perceptions. Originality/Value: This study innovatively applies AI to architectural design, offering objective support for experiential design and enhancing the understanding of the relationship between space and emotion.