AI-based mentoring systems provide personalized and adaptive support to learners in higher education, utilizing artificial intelligence to facilitate individualized learning pathways. These systems offer real-time feedback, promote self-regulated learning, and generate data-driven insights that enhance academic performance and learner engagement. This systematic literature review (SLR) evaluates the current state of research on the integration of AI in educational mentoring over the past decade. A comprehensive screening process was conducted across five databases, with 50 key studies selected from an initial pool of 710 publications, following predefined inclusion and exclusion criteria. The findings reveal that AI integration in mentoring predominantly falls into two categories: technology-driven and hybrid models, where human mentors collaborate with AI systems to provide holistic support. The most frequently employed tools include Chatbots, Intelligent Tutoring Systems (ITS), and AI-based Recommender Systems, all of which provide personalized learning trajectories, real-time adaptive feedback, and tailored support. Although AI-based mentoring systems show considerable potential for scaling individualized educational support, significant challenges remain, particularly in relation to technological integration, the digital divide, and data privacy. Furthermore, while AI offers cognitive support, current systems lack emotional intelligence, necessitating human mentors to provide social and emotional guidance. The review suggests that future research should prioritize advancing AI systems' emotional intelligence, developing hybrid mentoring models, and establishing ethical frameworks to ensure transparent, fair, and trustworthy AI deployment in higher education.