Designing AI-Based Slide Evaluation Support Systems

Effective presentation slides necessitate a blend of instructional design expertise, structural clarity, and subject matter proficiency, all supported by evaluation and feedback. However, providing human feedback is expensive and time-consuming, with limited availability of feedback providers. We propose an AI-based approach for evaluating and generating feedback on presentation slides. Using a Design Science Research approach, we derived meta-requirements from practitioner interviews and online user reviews, formulating five design principles grounded in Feedback Intervention Theory. Implementing these principles in a prototype, we offer slide creators guided evaluation and actionable feedback. We subjected the prototype to proof-of-concept and proof-of-value evaluations. Thus, we contribute a thoroughly validated guideline, developed from both solicited and unsolicited feedback, towards a nascent design theory for AI Slide Evaluation-augmented Support Systems for multi-modal content. Our study offers theoretical insights and practical implications for enhancing digital content creation via AI-based feedback mechanisms.