What changed
Bayer Crop Science announced the development of the Agentic Decision Engine (ADE), a system that enables natural language interaction with machine learning workflows. ADE is built on Amazon Bedrock AgentCore and integrates with AWS Unified Operations to support its deployment and operation.
What this means in practice
ADE allows users to ask questions about their ML workflows without needing to navigate complex systems. The system uses a multi-agent architecture where a supervisor agent routes queries to specialist agents that retrieve context from relevant systems. This reduces the time required to diagnose issues, such as identifying why a phenotyping workflow failed, from over an hour to under 10 minutes. ADE runs on Amazon Bedrock AgentCore, which provides a managed runtime, IAM-scoped tool access, and a framework-agnostic design. The system also leverages AWS Unified Operations, which includes a Technical Account Manager (TAM) and a generative AI Domain Specialist Engineer (DSE) to support the platform's development and operation.
Limitations and availability
ADE is currently in its first phase, resolving about 80 percent of user questions without an engineer in the loop. The system runs 22 specialist agents in production, each focused on a specific domain. The architecture includes four layers: user experience, agent orchestration, specialist agents, and tools/data sources. The platform is designed for use by breeders, analysts, researchers, data scientists, and data engineers working with plant breeding data. The supplied announcement does not specify the exact release date or availability beyond its current phase.
