Accelerate scientific research with governed, agentic artificial intelligence (AI). Effective June 2, 2026, Microsoft made Microsoft Discovery generally available and released the Microsoft Discovery app in preview.
Summary
- Microsoft Discovery is now generally available for all organizations as a platform for building and governing agentic AI workflows in research and development (R&D)
- The Microsoft Discovery Engine drives a repeatable loop from evidence to hypotheses, execution, analysis, and the next iteration
- Teams create specialized agents, connect them to institutional knowledge, and orchestrate modeling, simulation, analysis, and validation tools
- Governance, reproducibility, and transparency are built in, with outputs that show confidence scoring and cited findings
- The new Microsoft Discovery app is available in preview as a local desktop experience for researchers, students, and academic labs
- The app downloads from the Microsoft Discovery GitHub and starts with a GitHub Copilot account
What is changing?
Microsoft moved Microsoft Discovery from private preview to general availability for all organizations. The platform now targets production R&D environments, with governance and transparency built in so workflows stay reproducible and outputs stay reviewable.
Microsoft also released the Microsoft Discovery app in preview. This local desktop experience lowers the barrier to early exploration, letting researchers and small teams start in their own working environment and later bring that work into the platform.
The Microsoft Discovery workspace welcome experience.
What is Microsoft Discovery?
Microsoft Discovery is an Azure-based platform for building and governing agentic AI workflows across scientific and engineering disciplines. It lets teams create and coordinate specialized agents, connect them to institutional knowledge and external scientific information, and orchestrate modeling, simulation, analysis, and validation tools. It is designed to work within existing R&D environments, not replace them, and keeps human judgment at the center of decisions.
Updated (2026-07-20) Licensing note: Microsoft Discovery is generally available for organizations and runs on Azure Pay-As-You-Go. Billing uses a new unit called User Messages: one User Message covers ten backend agent operations and costs 0.20 US dollars. The Microsoft Discovery app is a separate preview that runs locally and consumes GitHub Copilot Credits, so it requires a GitHub Copilot account. Preview features and pricing are subject to change. Pricing note: all prices are non-binding and for information purposes only; only a current, precise quote is binding.
The Microsoft Discovery Engine task creation and status overview.
Customer benefits at a glance
- Reach defensible R&D decisions faster through repeatable, evidence-driven exploration
- Trust outputs you can review, reproduce, and audit
- Keep proprietary knowledge connected and governed appropriately
- Lower the barrier for early-stage and academic research to start with AI
- Scale local experiments into enterprise R&D programs without rework
Key capabilities
- Microsoft Discovery Engine that moves work from evidence to hypotheses, execution, analysis, and iteration
- Specialized agents coordinated across modeling, simulation, analysis, and validation tools
- Connections to institutional knowledge and external scientific information
- Outputs with confidence scoring and cited research findings
- A local desktop app for literature exploration, hypothesis generation, scientific reasoning, and iterative experimentation
- A growing partner ecosystem contributing integrated tools and domain expertise
The Microsoft Discovery Engine output with confidence scoring and cited research findings.
Why Microsoft Discovery is relevant now
Many R&D teams are moving AI from pilots to production, where reproducibility and traceability matter. Microsoft Discovery embeds governance and review into the workflow, which reduces compliance and audit risk. The repeatable loop shortens the path from idea to validated result, which saves research time. Confidence scoring and cited findings increase clarity for expert review. A clear separation between AI reasoning and human judgment increases control over scientific decisions.
What organizations should do now
- Review which R&D workflows could benefit from governed agentic AI
- Identify the institutional knowledge and tools your agents would need to connect
- Validate governance, reproducibility, and audit requirements before scaling
- Enable a small team to test the Microsoft Discovery app preview on one project within 30 days
- Update your Microsoft licensing plan to reflect Microsoft Discovery and related AI services
- Monitor preview features, which are subject to change
Ready to align your Microsoft Discovery licensing?
Assess the licensing impact of Microsoft Discovery and related Microsoft AI services with the help of our experts. Contact us to plan a cost-efficient Microsoft licensing setup today.
Sources
For the announcement of Microsoft Discovery: General Availability and New App Preview for R&D, please visit: https://azure.microsoft.com/en-us/blog/announcing-microsoft-discovery-general-availability-and-microsoft-discovery-app-preview/.
For useful software licensing information on Microsoft products, please visit: https://www.schneider.im/software/microsoft/.