Do Design Keywords Matter for Generative AI? An Empirical Study of a Design Element-Based Prompt Framework

Generative AI is reshaping early-stage product design by enabling rapid generation of visual concepts from text prompts, yet prompt construction remains largely experience-driven. This study evaluates a structured prompt framework for office consumer products that integrates theory-informed keyword elicitation, KJ clustering, Analytic Hierarchy Process (AHP) weighting, and comparative evaluation. Using computer-mouse design, we compared AHP-guided prompts with non-AHP baseline prompts. In the primary study, AHP-guided prompts achieved higher AHP-weighted composite scores and showed a consistent ranking tendency across aesthetics, emotional resonance, and visual usability; these findings are interpreted descriptively. A robustness study using controlled seeds and mixed-effects models provided inferential support for higher ratings of AHP-guided prompts in aesthetics and visual usability only. The framework supports more structured, interpretable, and traceable prompt construction, offering practical implications for AI-assisted design and human–AI co-creation.