EMS ChatBot — Slack-Based RAG System

A sophisticated Slack-based chatbot leveraging Retrieval Augmented Generation (RAG) architecture to provide intelligent responses about company policies and payroll data. The system combines multiple vector databases and AI technologies to deliver contextual, accurate information through natural language queries.

Key Features

Dual knowledge base system (company policies + payroll data)
Custom text chunking with RecursiveCharacterTextSplitter
Multi-metric vector similarity search (cosine, euclidean, dot_product)
Real-time message processing with thread support
User authorization middleware for secure access
Batch processing for large dataset vectorization
Dynamic date query parsing and MongoDB filter generation
Optimized batch operations for 50+ documents simultaneously
Intelligent date extraction and query conversion for data searches
Role-based access control for enterprise use

Technical Architecture

Node with Express-style architecture
Slack Bolt Framework for real-time Slack integration
RAG pipeline implementation with LangChain

Highlights

Successfully processed and vectorized company policy documents and payroll datasets, implemented intelligent date extraction and query conversion, and built role-based access control for enterprise use.

Streamlined HR and payroll query resolution, reduced manual policy lookup time, and provided 24/7 automated support for company information.