A step-by-step tutorial on implementing the AI Planning Pattern using Java, MongoDB, LangChain4j, and Jakarta EE. The pattern breaks complex AI objectives into smaller deterministic tool calls using a Reason-Act-Observe loop. The example builds a travel itinerary assistant where an LLM queries MongoDB for real city and attraction data via annotated tool classes, rather than relying on training data alone. Covers domain modeling with Jakarta NoSQL/Data, exposing services as LLM tools, and building a JSF/PrimeFaces UI on GlassFish.
Table of contents
PrerequisitesStep 1: Generate the ProjectStep 2: Generate the domain classesStep 3: Defining the AI layer over MongoDB integrationStep 4: Showing the result with UIStep 5: Execute the applicationConclusion573 Impressions