Build a contextual chatbot application using Amazon Bedrock Knowledge Bases

This article explains how to build a contextual chatbot application using Amazon Bedrock Knowledge Bases, a serverless service for retrieval augmented generation (RAG) with large language models (LLMs).

Specifically, the article covers:

• An overview of retrieval augmented generation (RAG) and its benefits for contextual chatbots
• How Amazon Bedrock Knowledge Bases provides a managed solution for RAG, including data ingestion and text generation workflows
• The solution architecture for a chatbot application using Amazon Bedrock Knowledge Bases, API Gateway, Lambda, React, and other AWS services
• Steps to deploy the solution using AWS CDK and test the chatbot application
• Sample questions and responses from the knowledge base (Amazon shareholder letters)
• Instructions to clean up the deployed resources