Optimization of Prompt Engineering in the Development of Intelligent Tutor Chatbots for Personalized Learning

Abstract

The advancement of Large Language Models (LLMs) opens new opportunities in education, particularly through the development of intelligent tutor chatbots capable of delivering personalized and adaptive learning experiences. However, the effectiveness of such systems depends heavily on the quality of the prompts or instructions provided to the AI model. This study aims to identify the most relevant prompt engineering components for intelligent tutor chatbots, analyze their contributions to lrning personalization, and propose a prompt design framework applicable to the educational context in Indonesia. The method used is a Systematic Literature Review (SLR) with the PRISMA protocol, involving the systematic selection of 13 reputable articles from the SCOPUS database published between 2021 and 2026. The findings indicate that structured prompt designs encompassing elements of role definition, audience specification, feedback style, Chain-of-Thought techniques, and output format significantly enhance the chatbot's ability to generate personal, adaptive, and pedagogically valuable responses. Five dominant thematic clusters were also identified: (1) the impact and challenges of LLMs in education, (2) educational chatbot design, (3) AI-based learning personalization, (4) prompt engineering and LLM evaluation, and (5) the integration of GenAI tutors, AI literacy, and learning analytics. This study concludes that prompt engineering is a crucial pedagogical instrument that can transform LLMs from mere answer machines into intelligent tutors that foster critical thinking and student learning autonomy.