Answer: Use Agents for Amazon Bedrock with Amazon Bedrock knowledge bases to build the application.

## Explanation **Why Option B is correct:** 1. **Agents for Amazon Bedrock** allow you to build generative AI applications that can execute tasks and interact with users using natural language. 2. **Amazon Bedrock knowledge bases** provide a way to connect foundation models to company-specific data sources, enabling the AI to access and retrieve information from documents and databases. 3. **The specific requirements** involve: - Checking open customer claims - Identifying details for specific claims - Accessing documents for claims These requirements align perfectly with Agents for Amazon Bedrock combined with knowledge bases, as this solution can: - Use natural language to query claim information - Retrieve specific claim details from structured data - Access and reference claim documents stored in knowledge bases **Why other options are incorrect:** - **Option A (Agents for Amazon Bedrock with Amazon Fraud Detector):** Amazon Fraud Detector is specifically designed for fraud detection, not for general claim management and document access. - **Option C (Amazon Personalize with Amazon Bedrock knowledge bases):** Amazon Personalize is for building recommendation systems, not for creating conversational AI applications for claim management. - **Option D (Amazon SageMaker to build a new ML model):** While SageMaker can build ML models, it would require significant development effort and wouldn't provide the ready-made conversational AI capabilities that Agents for Amazon Bedrock offers. This solution leverages AWS's managed generative AI services to create an application that can understand natural language queries, access claim data, and retrieve documents efficiently.

Author: Ritesh Yadav