Announcing Microsoft Discovery general availability and Microsoft Discovery app preview

Microsoft has released Microsoft Discovery as a generally available platform, marking a significant shift in how organizations can build and manage agentic AI workflows. If you’ve been following the AI space, you know that autonomous agents—AI systems that can plan, execute, and adapt without constant human intervention—are becoming increasingly central to enterprise automation. Microsoft Discovery is designed to solve one of the biggest challenges teams face: how to actually build these agents responsibly and at scale, not just experiment with them in isolated prototypes.

At its core, Microsoft Discovery provides a unified platform that brings together the tools you need to create agentic workflows while maintaining governance controls. Technically, it integrates with Azure’s AI services, allowing you to connect language models (like GPT-4), your existing APIs, and data sources into coordinated agent systems. Think of it as a workflow orchestration layer specifically built for AI agents—similar to how you might use a message queue or state machine to coordinate microservices, but with capabilities designed for agent decision-making and tool use. The platform includes monitoring and policy enforcement, which means you can define rules about what agents can access and do, then monitor their behavior in production. This is critical because unlike traditional automation where you hardcode every decision point, agents make contextual decisions that need oversight.

Where this gets practically useful is in scenarios like customer support automation, IT operations, and document processing. Imagine your support team using an agent that can independently triage incoming tickets, pull relevant customer history from your CRM, check your knowledge base, and route issues to the right specialist—all while maintaining audit trails that your compliance team can review. Or consider IT operations where an agent monitors infrastructure logs, investigates anomalies, performs diagnostics, and escalates appropriately without every alert needing manual review. Organizations testing these patterns have reported significant time savings, but they’ve also learned that unmonitored agents create risk. Microsoft Discovery’s governance layer addresses this directly, letting you set boundaries on agent autonomy while still capturing the efficiency gains.

The availability of Microsoft Discovery signals that the industry considers agent automation ready for mainstream enterprise adoption. For teams building on AWS or other platforms, this doesn’t necessarily mean you need to switch clouds—many of Microsoft’s agent capabilities integrate with multi-cloud setups. But it’s worth evaluating whether your current workflow automation stack has the agent-specific features you’ll need: built-in orchestration for multi-step agent reasoning, easy integration with your model choice, and native governance controls. If your team is currently stitching together Lambda functions, Step Functions, and custom monitoring to handle agent-like behavior, a dedicated platform might save you considerable engineering time.