Comparing Agentic AI Frameworks for Java
Timo Salm - Broadcom / Sandra Ahlgrimm - Microsoft
Agentic AI marks a shift from passive, prompt-driven interactions to autonomous systems that can reason, plan, and execute work. On the JVM, several frameworks are emerging to support this new paradigm, but they take very different approaches.
In this session, we compare frameworks such as Spring AI, LangChain4j, and Embabel on how they implement core agent capabilities such as reasoning loops, tool invocation, memory, and orchestration. We also look at how the Model Context Protocol (MCP) helps standardize access to data and tools across agents.
Through practical examples, attendees will gain a clear understanding of the trade-offs between these frameworks and guidance on choosing the right one to evolve JVM applications from simple LLM integrations into truly autonomous, agentic systems.