This systematic literature review investigates the landscape of Artificial Intelligence (AI) in mathematics education, emphasizing successful implementation models, positive learning outcomes, and key factors influencing effective adoption. Through a synthesis of recent empirical studies, the review identifies innovative applications of AI, including personalized tutoring platforms, AI-powered chatbots in e-textbooks, and adaptive learning systems designed for exam preparation. While acknowledging ethical challenges, such as potential biases and limitations in assessment capabilities, this research highlights Capacity of AI to enhance student engagement, foster deeper understanding of concepts, improve test preparedness, and offer more accessible learning experiences. The review also explores key pedagogical considerations and the importance of teacher collaboration, demonstrating that AI can be most effective when integrated into a framework that promotes ethical and equitable access. In alignment with the National Education Policy (NEP) 2020, this review advocates for a proactive approach that prioritizes teacher training, evidence-based practices, and technology implementation that strengthens existing pedagogies. Recommendations are provided for educators, policymakers, and researchers, underscoring the need for a collaborative ecosystem that leverages AI to transform mathematics education to meet the needs of all students while adhering to the pillars of the NEP 2020 and preparing learners for future challenges.