Game-inspired Pedagogical Conversational Agents – Design Knowledge and Empirical Studies

Driven by technological advancements, digital learning formats are becoming an integral part of how students learn. Pedagogical conversational agents (PCAs) have emerged as the leading trend in digital learning. These are artificial intelligence-based chatbots or voice assistants that support multiple students individually and independently of time and location. Despite their potential, research has rarely considered how to design these systems in a way that fosters motivation, thereby hindering the full realization of their value. Students often perceive learning as burdensome, which motivated this dissertation to investigate how to transform the learning process into a playful activity. To realize this vision, this work introduces “game-inspired PCAs” – conversational agents in educational contexts that are integrated with game-based approaches. This thesis distinguishes between two types: gamified PCAs, which embed individual game mechanics (e.g., points, quests) into the dialog, and PCAs in game-based learning environments, which operate within an existing application such as a serious game. Their design involves a complex interplay of motivational, educational, social, and technical aspects. This dissertation aimed to derive prescriptive design knowledge for game-inspired PCAs, to contribute to solving these design challenges, and to empirically investigate the motivational effect of corresponding instantiations. Based on the design science research (DSR) paradigm, design knowledge was developed and tested in the three-year practical project StudyBuddy; the results are documented in 44 peer-reviewed publications. Across seven main projects, we (my co-authors and I) engaged in continuous co-creation with learners to analyze challenges and needs, develop design prescriptions, and empirically validate the resulting prototypes. Through the interplay of the relevance cycle (practical relevance), rigor cycle (theoretical foundation), and design cycle, we developed game-inspired PCA artifacts for language learning, business information systems education, and creative entrepreneurship tasks and tested their motivational impact. We condensed the resulting design knowledge into design principles and discussed, classified, validated, and synthesized the insights in a final meta-study with the existing knowledge base. A newly introduced vision cycle anchored the vision of a playful, collaborative, and sustainable learning journey in core values such as positivity, collaboration, and the development of a resilient education system. Finally, we contributed five design principles (DPs) on goal-setting and reflection, novice-expert relationship, performance-related motivation, learning story narration, and challenging moderation of learning with peers. Each principle emphasizes a specific social role of the PCA in collaboration with learners, similar to different role expressions in classical games. We also formulated three meta-DPs regarding the playful framing of the learning journey as personal growth, the adaptive tailoring of game elements to individual needs, and the mindful handling of cognitive load. Thus, the dissertation structures the research field of PCAs and game-based learning and demonstrates their combination as an innovative approach that integrates collaborative dialog learning with motivational game approaches. By differentiating the design recommendations according to contexts (gamification and serious games, individual and group learning, low and high learning objectives), domain-specific boundary conditions in the implementation of game-inspired PCAs also become visible. Methodologically, the work extends DSR through the incorporation of a vision perspective and a meta-study approach, facilitating value-driven research efforts and the synthesis of design-theoretical knowledge. Practically, the dissertation offers advice for educators, developers, and educational institutions: Architectural templates, example prototypes, and guidance for making specific design decisions (e.g., choosing appropriate game mechanics) illustrate how to implement game-inspired PCAs in various areas. Overall, the work provides a framework upon which future researchers and practitioners in this dynamic field can build.