As Kubernetes adoption continues to grow, managing and optimizing storage systems in dynamic, containerized environments has become increasingly complex. This paper explores how AI-driven observability can revolutionize the monitoring and management of Kubernetes storage systems on AWS. We introduce a scalable framework that leverages machine learning models and analytics to detect anomalies, predict performance bottlenecks, and optimize storage resources in real-time. By integrating AWS-native services such as Amazon SageMaker, Amazon CloudWatch, and AWS OpenTelemetry with Kubernetes observability tools like Prometheus and Grafana, this solution empowers operators to achieve deeper insights and proactive control over their storage infrastructure.