Conventional cloud solutions continue to encounter challenges due to escalating resource capacity, an expanding array of features, and support for a multitude of business services. Leveraging the capabilities of AI-agents built on large language models, this study hypothesizes that they can emulate the functions of cloud components and collaborate for effective management of clouds. Our prototype framework utilizes multiple AI-agents, each responsible for specific tasks while supported by other components to ensure system integrity. Initial simulation results indicate a substantial reduction in service-level agreement violation and balanced server resource usage compared to widely used benchmarking algorithms.