Explore practical strategies for building and scaling AI agents in production environments through this 19-minute conference talk. Learn how to transition from local Model Control Protocol (MCP) implementations to robust cloud-scale architectures that engineering teams use to develop sophisticated agent systems. Discover real-world use cases through live demonstrations and gain insights into how product teams are shaping the future of agentic AI services. Understand the architectural patterns and engineering approaches necessary for deploying agents at enterprise scale, with practical examples and forward-looking perspectives on the evolution of agentic services in cloud environments.