AWS Shows How to Build a Natural-Language Claims Assistant With Amazon Bedrock Knowledge Bases
On September 30, 2026, AWS published a technical guide to building a conversational claims assistant that answers natural-language questions with citations. It covers ingesting claims documents from Amazon S3, querying through the AgenticRetrieveStream API, multi-turn follow-ups, metadata filters, and contextual grounding guardrails.

Full Report
AWS published the technical guide “Query claims in natural language with Amazon Bedrock Knowledge Bases” on September 30, 2026. It explains how to build a conversational claims assistant with Amazon Bedrock Knowledge Bases, including answers to natural-language questions with citations that connect responses to relevant source material.
The walkthrough covers ingesting claims documents from Amazon S3 and submitting queries through the AgenticRetrieveStream API. It also addresses follow-up questions across multiple turns, showing how users can continue an inquiry within the same conversation.
For query controls, the guide covers metadata filters and contextual grounding guardrails. The example therefore spans not only the question-and-answer interaction, but also retrieval conditions and answer grounding. A day earlier, AWS published a separate technical guide to natural-language questions about video, using a Strands Agents SDK agent to coordinate Amazon Bedrock, Amazon Rekognition, and Amazon Transcribe. The two guides address different data and architectures, while both show conversational questions connecting with AWS AI services.
Why This Matters
For claims-document question answering, the practical details extend beyond natural-language input: cited answers, metadata filtering, multi-turn follow-ups, and grounding controls all affect how the workflow is implemented. The guide brings those pieces together in one example, making it directly relevant to teams evaluating conversational retrieval for claims documents.
Evidence
Start with the primary evidence, then review supporting sources and context.
Primary Evidence
1Query claims in natural language with Amazon Bedrock Knowledge Bases
Shreya Pawaskar
“This technical how-to builds a conversational claims assistant on Amazon Bedrock Knowledge Bases that answers natural-language questions with citations. It covers ingesting claim documents from Amazon S3, querying with the AgenticRetrieveStream API, multi-turn follow-ups, metadata filters, and contextual grounding guardrails.”
Recommended signals
Recommended by brand, topic, and recent related coverage