Query

This demo shows a Retrieval-Augmented Generation (RAG) workflow built around a local LLM. Questions can be sent directly to the model, or routed through a retrieval step that searches a vector database built from the HR, Policy, and Sales document corpora. The most relevant document chunks are injected into the model prompt so answers are grounded in source material and returned with citations. The system also demonstrates simple automation logic by detecting when responses should be routed to Slack or Discord, with manual review safeguards for ambiguous cases.