Automating Workflows with Multi-Agent Systems, Kristiyan Stoyanov

This talk examines the limitations of using a single large language model (LLM) agent for complex workflows and highlights the advantages of multi-agent architectures as a scalable and controllable solution. The speaker introduces the concept of agents, the role of agent-to-agent (A2A) collaboration, and demonstrates how workflows can be structured as coordinated interactions among multiple agents. Using Strands as a conceptual framework, the session includes a live demonstration showcasing how various agents work together to complete a real task. Attendees gain practical insights into the functionality of multi-agent systems and their integration into backend architectures.