Microsoft Discovery Now Generally Available for Agentic AI Workflows in Science and Engineering

Microsoft took the stage at Microsoft Build today to announce that Microsoft Discovery is now generally available for all organizations. The platform offers a comprehensive environment for building and governing agentic AI workflows across scientific and engineering disciplines, marking a significant step in weaving AI deeper into the fabric of research and development.

What is Microsoft Discovery?

Microsoft Discovery is built to speed up scientific and engineering innovation. It lets researchers and engineers deploy AI agents that autonomously explore hypotheses, crunch data, and generate insights. The platform covers the full lifecycle of agentic workflows, from authoring and testing to deployment and governance, with built-in compliance and safety guardrails baked in from the start.

The general availability release integrates with Azure AI, Azure Machine Learning, and other Microsoft cloud services, allowing teams to scale their research while maintaining reproducibility and security. Organizations can now provision Discovery workspaces through the Azure portal and configure custom agent behaviors tailored to specific domain tasks.

The Microsoft Discovery App Preview

Alongside the GA announcement, Microsoft offered a first look at the Microsoft Discovery app, a new interface that brings agentic AI capabilities directly to domain experts. The app is designed to lower the barrier to entry for scientists and engineers who may not have deep coding chops, letting them interact with AI agents through natural language and visual dashboards.

Early testers have already put the app to work on tasks like planning chemical syntheses, discovering new materials, and optimizing experimental parameters. The app also includes collaborative features, enabling teams to share agent workflows and results in real time.

Governance and Compliance

What sets Microsoft Discovery apart is its governance framework. The platform includes role-based access controls, audit logging, and integration with Microsoft Purview for tracking data lineage. That makes it a natural fit for regulated industries such as pharmaceuticals, aerospace, and energy, where traceability and compliance are non-negotiable.

Microsoft stressed that agentic AI governance is a top priority, with the platform enforcing operational guardrails that prevent agents from executing unauthorized actions or accessing sensitive data without approval.

Integration with Agentic Ecosystems

Microsoft Discovery plugs into Copilot Studio and Azure AI Agent Service, letting organizations chain custom agents into larger workflows. For example, a chemistry agent could request data from a materials science agent, verify results through a simulation agent, and produce a final report with citations, all under human supervision.

The move is part of Microsoft's broader push to embed agentic AI into enterprise workflows, following earlier releases of Copilot agents and the Azure AI Agent Service.

Availability and Pricing

Microsoft Discovery is available starting today through the Azure Marketplace, with pay-as-you-go pricing based on the compute and storage resources used by agent workflows. The app preview is available for selected enterprise customers, with a broader rollout expected later this year.

To help early adopters hit the ground running, Microsoft has also published a set of reference architectures and example agent templates on GitHub.