RAG Assistant for Company Documents and Policies

Alessandro

Alessandro Federici

RAG Assistant for Company Documents / Policies

๐ŸŽฅ Video LOOM Overview (ENG ๐Ÿ‡ฌ๐Ÿ‡ง): https://www.loom.com/share/ae3630cfc13145f2b7f74bab540caa36 ๐ŸŽฅ Video LOOM Overview (ITA ๐Ÿ‡ฎ๐Ÿ‡น): https://www.loom.com/share/5dcbb84b6ab44e8a9e27f40a4a4837a7

This automation acts as a virtual company secretary, allowing employees to instantly ask questions about internal procedures or company policies directly through Slack or WhatsApp. Using a Retrieval-Augmented Generation (RAG) workflow, the system accesses and interprets company documents stored in Google Drive โ€” enabling real-time answers like:
โ€œHow do I report a business trip?โ€ โ€œWhatโ€™s the sick leave procedure?โ€
Whenever new documents are added to the companyโ€™s Drive folder, they are automatically processed, vectorized, and added to a Pinecone database for contextual search. From there, the chatbot can understand and respond based on the most relevant internal files, keeping company knowledge centralized and easily accessible for all employees.

Tools Used / Stack

n8n โ€“ Workflow automation and orchestration
Google Drive โ€“ Document storage and trigger for new uploads
OpenAI API โ€“ Embeddings + Chat Model for RAG and conversational logic
Pinecone โ€“ Vector database to store and retrieve document embeddings
Slack API โ€“ Employee communication interface
WhatsApp Business API โ€“ Alternate chat interface for employees
Webhook & HTTP modules โ€“ Custom data handling and routing logic

Key Features

๐Ÿ” Auto-Sync Documents โ€“ New or updated company files in Google Drive are automatically processed and indexed for semantic search.
๐Ÿ’ฌ Multi-Channel Interaction โ€“ Employees can ask questions either via Slack or WhatsApp; the assistant replies on the same platform.
๐Ÿงฉ RAG-Based Contextual Responses โ€“ Combines OpenAIโ€™s language model with Pinecone vector search for accurate, context-aware answers.
๐Ÿง  Persistent Memory โ€“ Simple Memory node ensures the AI agent maintains context during the conversation.
โš™๏ธ Dynamic Data Flow โ€“ The automation intelligently merges incoming data, adapts payloads, and routes responses depending on the trigger source.
๐Ÿ”’ Scalable & Private โ€“ Designed to operate entirely within the companyโ€™s private cloud structure, maintaining data privacy and security.

Outcome

This automation transforms internal knowledge management into an always-on AI assistant, dramatically reducing the workload for HR and administrative teams.
Internal inquiries and emails are reduced by over 90%, limited to exceptional cases or procedures not present in company documents.
The result is significant time savings, greater operational efficiency, and truly accessible company knowledge, transforming internal policy documents into an intelligent, conversational knowledge base.
Drive folder to be monitored for new business documents
Drive folder to be monitored for new business documents
Question in English or any other language will be automatically translated
Question in English or any other language will be automatically translated
Answers for the previous questions will be sent in DM / private for privacy
Answers for the previous questions will be sent in DM / private for privacy
Chat with Assistant on Whatsapp in any language
Chat with Assistant on Whatsapp in any language
Example of Business Policy Document PDF
Example of Business Policy Document PDF
Like this project

Posted Nov 9, 2025

Developed a RAG-based virtual assistant for company document queries via Slack and WhatsApp.