Building generative AI applications in the cloud can feel overwhelming with the vast number of tools, services, and patterns available. In this talk, we’ll cut through the noise and explore a decision-making framework to choose the right stack for your GenAI use case on AWS. We’ll compare options such as Bedrock Agents, SDKs like Vercel AI and Strands, and emerging architectures like Multi-Agent Control Plane (MCP) and Agent-to-Agent (A2A). Through real-world examples and lessons learned, you’ll gain practical insights to avoid common pitfalls and design scalable, observable, and context-aware GenAI solutions. Whether you’re starting your first prototype or optimizing for production, this session will help you turn complexity into clarity.