In this AWS re:Invent 2023 session, Victoria and Linda from AWS Developer Relations demonstrate how to unlock data insights using generative AI services. They showcase two approaches for retrieval-augmented generation (RAG): one using Amazon SageMaker and Amazon Kendra, and another utilizing Amazon Bedrock and vector databases. The presenters highlight the importance of customizing foundation models with enterprise data to mitigate hallucinations and improve accuracy. They provide live coding demos, showcasing the power of Amazon CodeWhisperer for efficient development. The session covers key AWS services like SageMaker, Kendra, Bedrock, and CodeWhisperer, emphasizing their roles in building powerful, customized AI applications. Attendees learn about different data preparation techniques, model deployment options, and the benefits of using vector databases for similarity search. The presenters stress the importance of choosing the right approach based on specific use cases and data types.