Building a triadic model of technology, motivation, and engagement: a mixed-methods study of AI teaching assistants in design theory education

In design education, it is often more difficult to keep students engaged in theory courses than in hands-on studio classes. Theory courses focus on abstract concepts like design history and principles, which can feel disconnected from practical experience. This study explores how AI-powered teaching assistants can support student engagement in design theory through a mixed-methods approach. Based on Self-Determination Theory (SDT) and Task-Technology Fit (TTF) Theory, we developed a triadic engagement model and tested it with data from 363 undergraduate design students who used a domain-specific AI assistant. Results from Partial Least Squares Structural Equation Modeling (PLS-SEM) and Artificial Neural Networks (ANN) show that communication quality, perceived competence, task-technology fit, and school support are key predictors of engagement. In contrast, individual technology fit and lecturer support have limited effects. Fuzzy-set Qualitative Comparative Analysis (fsQCA) identifies five learner profiles leading to high engagement, showing that different combinations of motivation, support, and technology fit can be effective. Interviews with 10 students identify three themes, further revealing that while the AI assistant is helpful and accessible, it lacks depth in critical thinking, and it demonstrates that students learn to verify AI assistants' responses and reflect on their learning. This study contributes to education and AI research by showing that chatbots must support both psychological needs and task alignment to foster meaningful engagement. It positions AI not just as an information tool, but as a partner in reflective and autonomous learning.