This chapter examines the evolving landscape of educational chatbots with a particular emphasis on personalized learning. While previous adaptations have primarily concentrated on content tailored to students’ performance, this text focuses on adapting learning styles, drawing upon the framework proposed by Kolb. The study employs large language models (LLM), instructed to generate responses aligned with Kolb’s learning styles and demonstrates the feasibility of an adaptive conversational style that is tailored to individual learning preferences with reasonable precision. The findings indicate that LLMs can generate responses aligned with different learning styles when prompted. An analysis employing linguistic metrics revealed notable variations, including readability and lexical richness. Furthermore, a classifier was trained to distinguish between the different learning styles with an acceptable degree of accuracy. Additionally, human evaluation revealed acceptable alignment with the requested learning style. This paves the way for the development of personalized training systems that can adapt not only to individual content but also to individual learning styles.