Session Overview

Join me for an insider’s look at Alfred, Givaudan’s secure, in-house ChatGPT solution built from scratch. In this session, we reveal why we needed a robust, internal conversational assistant and how leveraging Microsoft Azure made it all possible.

Learn about our step-by-step journey—from navigating stringent security regulations and setting up the necessary Azure subscriptions, to replicating ChatGPT’s UI and addressing challenges like saving user messages and managing conversation context length.

We’ll also share our production deployment experience and preview exciting future enhancements, including image generation and RAG integration with a document viewer. Whether you’re a tech leader or an innovation enthusiast, discover how to transform an idea into an enterprise-grade AI assistant.

Watch the Session Video

Presentation Slides

Key Takeaways

• Learn why enterprises need secure, internal AI assistants
• Understand the architecture and security considerations for building ChatGPT-like solutions
• Navigate Azure setup and configuration for enterprise AI
• Address challenges in UI replication, message persistence, and context management
• Discover production deployment strategies and lessons learned
• Explore future enhancements: image generation and RAG integration

About Alfred

Alfred is Givaudan’s internal secure ChatGPT solution that demonstrates how to successfully implement an enterprise-grade conversational AI assistant while maintaining the highest security standards.

Resources

• Azure OpenAI Service Documentation
• Azure AI Services
• Enterprise AI Best Practices