Building AI Agents with Model Context Protocol: From Specification to Implementation

This conference talk explores the Model Context Protocol (MCP) as a standardized approach for AI agent development, focusing on how it enables AI models to interact with external tools and resources. Christian Tzolov demonstrates MCP's architecture and core capabilities including tool discovery and execution, resource management with URI templates, and dynamic prompt handling. Learn how the Spring AI MCP Java SDK implementation transforms these specifications into practical agent capabilities, allowing traditional AI applications to evolve into full-fledged agents that interact with external systems, manage resources, and handle complex workflows through standardized interfaces. Through live coding examples, discover how to combine the Spring AI framework with MCP to create agents that effectively understand context, make decisions, and take actions across various external systems.