LangChain4J Beyond The Hype: From Prompt to Productivity, Agentic AI in Practice
Details
Agentic AI is moving fast — but separating real productivity from “agent-washing” is now the developer’s challenge.
Citi has recently shared public progress in this space, including internal GenAI platforms such as Citi Assist, Citi Stylus / Stylus Workspaces, and updates introducing agentic AI capabilities inside Stylus Workspaces aimed at improving employee productivity.
In this session, we will delve into:
How do you implement Agentic AI in Java so it reliably delivers productivity — not just hype?
We’ll start with a simple LangChain4J example and evolve it step-by-step into a workflow that demonstrates:
-
how to go from a prompt → structured execution
-
when an agent is useful vs when it’s unnecessary complexity
-
tool calling & orchestration patterns that work in real apps
-
how to get repeatable outcomes instead of “it worked once”
This is not a futuristic vision talk — it’s a hands-on, developer-first walkthrough showing Agentic AI done properly in Java with LangChain4J.
This session will be an engaging discussion with focus on AI from a practical engineering perspective. Everyone welcome.
Closing Lightning Talk
Composable Intelligence
Designing AI Native Workflows with Skills, Skill Composition and Agent Collaboration
Niall McLoughlin (DailyPay)