Research on Simplified Design of Interface Interaction Process Based on TRIZ and Generative AI
Miao Li, Shengzhe Liu, Bin Wu, Xueru Zhao, Zihuan Ning · International Journal of Human-Computer Interaction · 2026
Digital applications frequently suffer from interaction overhead caused by functional coupling in user interfaces, leading to high cognitive load and complex interaction flows. This study proposes a GAI-TRIZ framework that integrates generative AI with TRIZ (theory of inventive problem solving) methodology under human-centered AI principles to systematically optimize interface interaction processes. Following a “Human-in-the-loop” paradigm, GAI handles automated process functional analysis and trimming candidate generation, while human experts retain validation authority over AI-generated proposals. We applied the framework to a representative e-commerce shopping flow and conducted a within-subjects user study (N = 12). Results demonstrate significant improvements: task completion time reduced by 31.2% (p < 0.01), interaction steps decreased by 42.1% (p < 0.01), and System Usability Scale scores improved from 58.3 to 76.8 (p < 0.05). These findings validate that the GAI-TRIZ framework effectively reduces procedural burden while maintaining user control through structured human–AI collaboration.