A leading institute specializing in food science consulting wanted their content to be accessible as part of a Retrieval-Augmented Generation (RAG) application designed to support scientific research and innovation. This institute creates a variety of journal publications and their own food science-related content (like journals, courses, and newsletters). The primary goal of the application was to assist food scientists in the development of new food products by enabling AI-Powered intelligent Search and interaction with their extensive library of scientific content. 

Our Approach 

We focused on data engineering and AI integration, leveraging the Microsoft Azure ecosystem to build a scalable and intelligent solution. 

• Content Storage & Processing: We used Azure Storage to house XML and PDF components of the institute’s content. PDF files were orchestrated and processed using Azure Databricks, while Azure AI Document Intelligence enabled us to extract structured data—including text, tables, images, and metadata—from scientific publications. 

• Metadata Enrichment: Extracted metadata such as publication dates, authorship, external links, and content type was linked across documents to enhance discoverability and context. 

• Video Accessibility: To make video content searchable and usable, we integrated Azure AI Video Indexer, allowing users to explore video-based insights alongside written materials. 

• AI-Powered Chatbot: All processed data was fed into a custom-built AI chatbot powered by Azure AI Search. This chatbot was trained on a vast corpus of food science material, enabling it to provide expert-level responses and facilitate meaningful conversations with users. 

The Impact 

The solution transformed how food scientists interact with information: 

• Expert-Led Conversations: Users can engage with the chatbot as if consulting a food science expert. By prompting it with phrases like “You are a food scientist. Analyze XYZ,” users receive in-depth, contextual responses that support research and discovery. 

• Accessible Knowledge: Scientists can explore complex topics from both expert and non-expert perspectives, helping them reframe problems, validate ideas, and uncover new insights. 

• Enhanced Decision-Making: The chatbot supports evidence-based decision-making by surfacing relevant, peer-reviewed content in real time. 

Final Thoughts 

This project exemplifies how AI can be used not to replace experts, but to augment their capabilities. By modernizing access to scientific content and enabling intelligent interaction, we’ve helped food scientists unlock new opportunities for innovation and collaboration. 

Our team has been working hard these last few months, and we are excited to share our success with you. Don’t forget to subscribe to our newsletter to stay updated on our most recent  Microsoft 365  projects.