[Workshop]
Implementing agents collaboration patterns with Spring AI and Bedrock Agentcore
Ruben Afonso - AWS / Yuriy Bezsonov - AWS / Arnaud Jean - AWS
As AI agents become increasingly integral to enterprise applications, multi-agent collaboration patterns are emerging as essential architectures for solving complex business problems. This hands-on 2-hour workshop explores building collaborative AI agent systems using Spring AI integrated with Amazon Bedrock AgentCore, with emphasis on observability, monitoring, and explainability across agent interactions.
Participants will learn how to design and implement production-ready collaborative agent systems that leverage multiple specialized agents working together to accomplish complex tasks. The workshop covers key collaboration patterns while emphasizing practical approaches to monitoring agent behavior and implementing explainability mechanisms that provide transparency into both individual agent decisions and collaborative workflows.
Key Learning Objectives:
- Design agent collaboration patterns using Spring AI and deploy them on Amazon Bedrock AgentCore.
- Build supervisor-worker, peer-to-peer, and hierarchical multi-agent architectures.
- Apply explainability techniques to understand and communicate agent reasoning and decision paths.
- Design dashboards and alerting mechanisms for proactive agent management.
- Leverage AWS native services such as Amazon CloudWatch and AWS Cost Explorer for comprehensive agent monitoring and TCO tracking.