Putting it all together: A customer support AI agent - Amazon Bedrock AgentCore

Amazon Bedrock AgentCore is in preview release and is subject to change.

Putting it all together: A customer support AI agent

Consider a customer, Sarah, who engages with your shopping website's support AI agent to inquire about a delayed order.

Memory AgentCore Memory

The interaction flow through the AgentCore Memory APIs would look like this:

  • Create a short term and long term memory and add strategies to create your long term memory.

  • Starting the session: When Sarah initiates the conversation, the agent creates a new, and unique, session ID to track this interaction separately.

  • Capturing conversation history: As Sarah explains her issue, each message (both her questions and agent's responses) is saved as an event using CreateEvent operation ensuring the full conversation is recorded in sequence.

  • Generating long-term memory: In the background, the asynchronous extraction process runs every few turns. This process analyzes the recent raw events using built-in or custom memory strategies (that you had configured when setting up AgentCore Memory through CreateMemory operation) to extract long-term memories such as summaries, semantic facts, or user preferences, which are then stored for future use.

  • Retrieving past interactions from short-term memory: To provide context-aware assistance, the agent calls ListEvents to load conversation histories. This helps the agent understand what issues Sarah has raised before.

  • Using long-term memories for personalized assistance: The agent calls RetrieveMemoryRecords, which performs a semantic search across extracted long-term memories to find relevant insights about Sarah's preferences, order history, or past concerns. This lets the agent provide highly personalized assistance without needing to ask Sarah to repeat information she has already shared in previous chats.

This integrated approach allows the agent to maintain rich context across sessions, recognize returning customers, recall important details, and deliver personalized experiences seamlessly, resulting in faster, more natural, and effective customer support.