Build memory-aware AI agents with Agentic Retrieve on Amazon Bedrock Managed Knowledge Base
Amazon Bedrock Managed Knowledge Base and Amazon Bedrock AgentCore memory now store conversations, extract facts, and bring long-term memory into a single AgenticRetrieveStream call. This post shows how to build a memory-aware, multi-user assistant where each user sees only the documents they may read and each user's memory stays private.
Amazon Bedrock Managed Knowledge Base and Amazon Bedrock AgentCore Memory enable users to build memory-aware AI agents. These agents can remember context from previous interactions, providing tailored responses grounded in relevant documents and user-specific information. Users expect their AI assistants to remember them and maintain context across turns, requiring applications to manage and utilize this context effectively.
Amazon Bedrock Managed Knowledge Base addresses this by integrating conversation history, fact extraction, and retrieval into its operations. When users submit questions, the system extracts pertinent facts from previous conversations and incorporates them into subsequent searches. This process happens automatically and independently for each user, ensuring privacy and context awareness. Additionally, the system grants users access to documents based on their role or permissions, maintaining security and relevance.
The integration of Amazon Bedrock AgentCore Memory into Amazon Bedrock Managed Knowledge Base streamlines the process of storing conversations, extracting facts, and integrating them into search queries. Applications can leverage this capability by simply adding a memoryConfiguration to the AgenticRetrieveStream request. Once enabled, the system autonomously retrieves relevant facts from memory and uses them to enhance search queries.
This integration ensures that the AI agent can maintain context and provide accurate, context-aware responses.
In the context of AnyCompany, an internal employee assistant, the system authenticates users via Amazon Cognito, using the ID token to determine the user's identity and permissions. The assistant then calls AgenticRetrieveStream with the necessary user context and memory settings, allowing it to retrieve and use relevant documents and long-term memory information.
This setup ensures that each user sees only the documents they are authorized to access, with their memory kept private and isolated. The system logs trace events for monitoring and debugging, providing insights into when and how the planner utilizes memory. By implementing these features, AnyCompany can create a personalized, context-aware assistant that enhances employee productivity while maintaining data security and privacy.
Written by urgent.news from AWS Machine Learning's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.