AWS is unveiling new multi-agent capabilities for its Amazon Bedrock platform, which will allow companies to build and manage complex workflows with multiple AI agents.
During his keynote at the AWS re:Invent conference, CEO Matt Garman stated that customers need agent collaboration to perform complex tasks, such as financial analysis, more efficiently, according to a report from VentureBeat.
Coordination of multiple agents
Garman explained that complex processes often require multiple specialized agents. “Creating a system that can coordinate multiple agents, share context across them, and dynamically route different tasks to the right agent requires specialized tools and generative AI expertise that many companies do not have available.”
With the new features in Bedrock, companies can build an orchestration agent to coordinate tasks. And to give agents access to needed information and enable parallel processes. Once tasks are completed, the orchestration agent collects the results.
For example, Moody’s, one of the first users, developed agents that analyzed macroeconomic trends and business risks, resulting in more accurate risk assessments.
AWS’s unique approach
AWS sets itself apart with a production-ready focus that allows companies to move from prototype to deployment faster. Orchestration features streamline workflows, manage status sharing, and dynamically assign tasks to specialized agents. This sets AWS apart from competitors such as Microsoft, which offers broad frameworks but places less emphasis on orchestration.
According to AWS, multi-agent collaboration allows companies to assign specific agents to project steps and speed up tasks by having multiple agents working in parallel. Companies build specialized agents on Bedrock and add a supervisor or orchestration agent that distributes tasks, shares information, and coordinates parallel actions.
Multi-agents: a growing standard
AI agents are becoming increasingly popular for simplifying workflows. Competitors such as Microsoft, ServiceNow, and OpenAI offer similar solutions, including agent libraries and orchestration tools. However, AWS is targeting enterprises that want to build their agent ecosystems, including a robust orchestration layer to control the proliferation of agents.