Investment advisory services are increasingly exposed to sophisticated fraud schemes and dynamic regulatory environments, necessitating advanced technological solutions for risk mitigation and compliance assurance. Machine learning (ML), with its capacity for real-time pattern recognition, anomaly detection, and predictive analytics, has emerged as a transformative tool in this sector. This review explores the landscape of ML-powered systems specifically designed for fraud prevention and compliance forecasting in investment advisory domains. It examines supervised and unsupervised learning approaches for detecting fraudulent activities, evaluates regulatory compliance models driven by natural language processing (NLP), and highlights the integration of real-time risk engines. Furthermore, the paper assesses challenges such as data privacy, model interpretability, regulatory alignment, and the implications of algorithmic bias. By synthesizing recent advancements and practical applications, this review underscores the role of ML in enhancing financial integrity, regulatory readiness, and operational resilience across advisory practices. The study concludes with strategic recommendations for future research and deployment within a governance-driven AI framework.