Gamification Empowered by LLM-Based Agents: A Systematic Literature Review

Abstract

The combination of gamification and agent systems based on Large Language Models (LLMs) presents a great opportunity in the design of adaptable, interactive systems that offer engagement and personalized interactions. In this systematic literature review, we examine the state of the art at this intersection, drawing on 11 studies published from 2019 to 2025, using the PRISMA 2020 framework, and investigating the following seven research questions related to: publication trends, areas of application, agent roles, gamification strategies, design goals, and limitations. We find LLM agents are being deployed across multiple application areas in gamified systems, in the fields of education, mental health, cyber defense, and commerce, while functioning in multiple roles such as tutors, an emotional companions, and being involved in the co-design process. Yet, we identified key challenges including weak connectivity between gamification mechanics and agents' behaviors and capabilities, limited longitudinal evaluation, and ethical concerns for LLMs related to persuasion and autonomy of the user. Our review provides a foundation for furthering the study and development of gamification systems involving LLM agents.