How Audi improved their chat experience with Generative AI on Amazon SageMaker

The article discusses how Audi improved their enterprise search experience using a Generative AI chatbot solution on Amazon SageMaker. The chatbot uses Retrieval Augmented Generation (RAG) technique, leveraging services like Amazon SageMaker, Amazon OpenSearch Service, and others.

Specifically, the article covers:

• The challenges Audi faced with their growing internal knowledge base and documentation
• Solution overview - data ingestion from sources like Confluence, and chatbot inference using RAG
• Key features - security, reduced hallucination, low latency, and integration with multiple sources
• Implementation details - use of LLMs like Falcon 7B-Instruct and Llama-2 13B-Chat, Amazon OpenSearch as vector database, Lambda functions, VPC, and other AWS services
• Results and comparison of LLMs, and experimentation on Audi documentation
• Conclusion on improving search experience and future enhancements