Introducing guardrails in Amazon Bedrock Knowledge Bases

This article introduces the integration of guardrails with Knowledge Bases for Amazon Bedrock, a service that helps connect foundation models to company data using Retrieval Augmented Generation (RAG). It provides enhanced safety and compliance for generative AI RAG applications by allowing you to filter harmful content and protect sensitive information in your documents.

Specifically, the article covers:

• Solution overview and workflow of applying guardrails to Knowledge Bases
• Use cases like legal knowledge management, financial services, and ecommerce customer support
• Steps to prepare a sample dataset, create a knowledge base, and sync data
• How to create guardrails with denied topics and sensitive information filters
• Testing the knowledge base with and without guardrails on the AWS console
• Code sample for querying the knowledge base with guardrails using AWS SDK
• Cleanup steps to delete the resources created