Large Language Models (LLMs) have emerged as a powerful technology with the potential to revolutionize education. This study investigates the adoption of LLMs in the University of Sofia “St. Kliment Ohridski” and the Faculty of Mathematics and Informatics with a focus on the factors influencing their implementation. A mixed-methods research design was employed to collect both quantitative and qualitative data. The findings indicate that educators perceive LLMs as valuable tools for personalized learning, automated tasks, and content creation. However, challenges such as lack of training, technical difficulties, and data privacy concerns hinder their widespread adoption. To maximize the benefits of LLMs, it is crucial to address these challenges through targeted interventions, including professional development, technical support, and clear guidelines for ethical use. Future research should explore the long-term impact of LLMs on student learning outcomes and educational equity.