While lots of work in AI agentic applications is happening in experimental projects, moving them to production in enterprise environments presents unique challenges. This talk will demonstrates how Spring AI leverages Java's core strengths - modularity, testability, and enterprise connectivity - to create production-ready agentic applications that integrate seamlessly with existing business services.

The talk will explore a practical implementations showing how to:

• Build robust, testable AI agents using Spring AI and AWS Bedrock
• Expose existing Java services using Model Context Protocol (MCP) with just a few lines of code.
• Combine techniques including conversation memory, RAG, tool calling and multi-agent workflows to develop rich agentic applications.