Agent Requirements

Agent Objective:

The goal of this RAG agent is to improve the llm context and add company information into is knowledge for better answers

Autonomy & Constraints:

It must be able to retrieve proper informations from the documents by itself

Tool & Action Design

The agent have a single tool, a vector DB were he can query the documents

Agent Architecture

This is a single agent architecture. The agent will have access to the retrieval tool and be able to query it in case it needs more informations to anwser

Memory & Context Design

The short term memory in handled by the chat memory database, which stores the conversations.

The Vector stores allows RAG sear