Made In The Shade

VND developed a secure, private AI chatbot that helps internal teams search and interact with extensive vendor documentation, including product manuals, specifications, pricing guides, tables and other technical PDFs. Staff ask questions in natural language and receive answers drawn from approved company sources, with references back to the underlying documents so users can see where information originated.

The solution uses the OpenAI API for language understanding and response generation, combined with a Retrieval-Augmented Generation (RAG) architecture. Rather than relying on the model general knowledge, the system retrieves relevant content from the client private document repository and supplies it as context for each answer.

How it works

• Document management. Approved vendor documents are maintained in the client SharePoint environment and synchronized to a private VND environment through the Microsoft Graph API.
• Document processing. Azure AI Document Intelligence extracts text and tables from complex and image-based PDFs and separates them into sections ready for indexing.
• Vector indexing. Processed content is converted to vector representations and stored in Qdrant, so search follows meaning and context instead of exact keywords.
• Intelligent retrieval. Each question is analyzed to identify the relevant vendor and category before the appropriate part of the knowledge base is searched.
• Response generation. Retrieved content is passed to OpenAI as context, and the answer is generated from the client documentation with source references.

Security and privacy

The assistant runs in a controlled, private environment. Client documents, vector indexes and application logic stay within controlled infrastructure, access is limited to authorized users through Microsoft 365 and Windows authentication, and the OpenAI API is used only for model inference. New vendor documentation can be processed, indexed and made available to the chatbot without retraining the underlying model.

Technology

OpenAI API, RAG, Qdrant vector database, Azure AI Document Intelligence, Python, PHP, WordPress, Microsoft Graph API, Microsoft 365 authentication, Docker and Linux.

Looking at AI for your own business? Explore our AI services or start with a free technology assessment.

Private AI for your own data

The same architecture applies to any organization that holds a large body of private documents and needs its people to find answers quickly. Healthcare providers, law firms, financial services and professional services teams can each run a private assistant grounded only in their own approved records, policies and reference material, hosted in a controlled environment. See how we approach this on our AI Implementation and AI Agents and Custom Development pages, or start with a free technology assessment.