Back to all apps
Cellenza Internal Assistant illustration
RAGChatbotAzure AI SearchAzure OpenAITeams

Cellenza Internal Assistant

Enterprise knowledge-base chatbot with RAG architecture, giving Cellenza employees instant answers from internal documentation, policies and project archives.

Author: Marie DupontPublished: 10 December 2024

About this app

Overview

The Cellenza Internal Assistant is a Retrieval-Augmented Generation (RAG) chatbot that indexes all internal documentation—HR policies, technical guides, project archives, and Confluence pages—and makes them instantly queryable via a natural-language chat interface.

Key Features

  • Semantic search: Azure AI Search with vector embeddings for accurate retrieval
  • Source citations: Every answer cites the source documents
  • Role-based access: Respects SharePoint permission groups
  • Conversation history: Multi-turn conversations with memory
  • Feedback loop: Users rate answers to improve the system over time
  • Teams integration: Available as a Microsoft Teams bot

Technical Stack

  • Frontend: React (Teams Tab + standalone web app)
  • Backend: Python FastAPI
  • Vector DB: Azure AI Search (with semantic ranking)
  • Embeddings: Azure OpenAI text-embedding-3-large
  • LLM: Azure OpenAI GPT-4o
  • Auth: Azure Entra ID (SSO)
  • Indexing: Azure AI Document Intelligence + LangChain

Data Sources Indexed

  • Confluence knowledge base
  • SharePoint document libraries
  • HR policy portal
  • Technical runbooks
  • Previous project deliverables

Privacy & Security

All data remains within the Cellenza Azure tenant. No data is sent to third-party AI providers — uses private Azure OpenAI endpoints.