This session explores how AIOps can be built by analyzing large-scale cloud-native operational data and applying AI-driven decision-making. Focusing on Kubernetes environments, it demonstrates how multi-agent architectures powered by Amazon Bedrock AgentCore and K8sGPT can automate fault detection, root cause analysis, and remediation. By combining domain expertise, observability data, and AI planning mechanisms, the solution reduces manual troubleshooting, shortens MTTR, and lowers operational complexity. The session also showcases voice-driven operations and human-in-the-loop safeguards, illustrating how intelligent automation can deliver reliable, auditable, and scalable cloud operations.