Uploaded May 2026 | Updated September 2026, 2 weeks ago

---

Most AI chatbot demos look smart… until you ask a follow-up question.

In this video, I build a Spring Boot RAG chatbot that answers questions using policy data retrieved from Oracle AI Vector Search.

Instead of letting the model guess, Spring AI retrieves the relevant policy documents from Oracle first and adds them to the prompt before generation.

The result is a chatbot with grounded answers — but still no memory.

That’s the point.

This is the first video in a series where I’m building a real AI application step by step:

* retrieval
* memory
* actions
* workflows

Repo:
[github.com/markxnelson/shopassist](https://github.com/markxnelson/shopassist)

#springboot #springai #oracle #rag #java #ai