GenAI Grows Up - Building Production-Ready Agents on the JVM

Explore a comprehensive conference talk that addresses the critical challenges facing GenAI implementation in enterprise environments and presents a JVM-based solution for building production-ready AI agents. Discover why most GenAI projects fail despite the technology's potential, examining common pitfalls including treating GenAI as standalone technology, excessive reliance on prompt engineering, integration problems with existing systems, and organizational issues stemming from top-down mandates. Learn how to overcome nondeterminism in AI systems through structured approaches and domain-integrated context engineering that leverages existing business logic and data. Understand the advantages of building AI agents on the JVM platform, including seamless integration with enterprise systems, mature tooling, battle-tested reliability, and access to critical business infrastructure that most enterprises already run. Get introduced to the Embabel agent framework, which combines GenAI capabilities with JVM strengths, featuring modern Java/Kotlin APIs built on Spring framework for creating testable, maintainable, and production-ready agents. Examine practical examples including a bank support agent implementation and see how the framework enables seamless code actions and LLM invocation while maintaining software engineering best practices like domain modeling, unit testing, and type safety. Compare JVM-based approaches with Python frameworks and understand why the Java community is positioned to lead in enterprise AI development, moving beyond imitation of Python approaches to build superior solutions for business-critical applications.