In this article How Microsoft Discovery supports R&D workflows at scale Expanding access with the Microsoft Discovery app preview Applying Microsoft Discovery across R&D Breakthroughs in science and engineering rarely come from a single insight. They emerge through cycles of hypothesis, experimentation, refinement, and review across teams, tools, and data. Today at Microsoft Build, we are announcing that Microsoft Discovery is now generally available for all organizations, providing a comprehensive platform for building and governing agentic AI workflows across scientific and engineering disciplines. We are also introducing the Microsoft Discovery app in preview, a local desktop experience that helps researchers, students, and scientific teams begin working with Microsoft Discovery today. Explore Microsoft Discovery features Since introducing Microsoft Discovery in private preview at Microsoft Build last year , we have worked closely with organizations applying AI to complex research and development (R&D) workflows. Their feedback helped reinforce where agentic AI needs to go beyond individual assistance, like supporting the iterative loops, evidence preservation, and tool coordination that define scientific work. The most challenging problems in R&D require more than just a prompt interface or a single model response. Scientific workflows require: Integration with institutional knowledge and domain expertise. Access to specialized modeling, simulation, and analysis tools. Connection to experimental evidence and validation data. Support for review processes that shape research decisions. A materials scientist may need to evaluate performance, safety, and cost alongside manufacturability and regulatory constraints. A semiconductor team may need to explore a larger design space without losing physical fidelity or traceability. A life sciences researcher may need to connect literature and experimental data with models and cohort-level evidence before deciding what to validate next. Microsoft Discovery is designed to work within these existing R&D environments , not replace them. The platform helps experts understand the reasoning path behind outputs and keeps human judgment at the center of scientific and engineering decisions. The general availability of Microsoft Discovery marks a significant milestone in turning these requirements into a production-ready platform for R&D environments with governance and transparency built in. Figure 1: The Microsoft Discovery workspace welcome experience . How Microsoft Discovery supports R&D workflows at scale Microsoft Discovery enables organizations to define agentic workflows around their own R&D programs. Teams can create and coordinate specialized agents, connect those agents to institutional knowledge and external scientific information, and orchestrate work across modeling, simulation, analysis, and validation tools. At the center of the platform is the Microsoft Discovery Engine , which supports the core loop of scientific work by helping teams move from evidence to hypotheses, through execution and analysis, and into the next iteration. This loop allows teams to move beyond isolated analysis toward repeatable, evidence-driven exploration, where they can compare tradeoffs, question assumptions, and narrow a search space in a way that can be reviewed and repeated. Figure 2: The Microsoft Discovery Engine task creation and status overview . As we continued product development, we focused on what it takes to bring agentic AI into production R&D environments : Workflows need to remain reproducible. Outputs must be reviewable. Proprietary knowledge must be connected and governed appropriately. Agentic systems need to fit into the operating model of R&D organizations. Those considerations, along with continued customer feedback, helped shape the general availability release and the platform capabilities behind it. Figure 3: The Microsoft Discovery Engine output with confidence scoring and cited research findings . Expanding access with the Microsoft Discovery app preview An important goal for Microsoft is to make advanced AI and computing capabilities more accessible to the people working on some of today’s most difficult scientific and engineering challenges. Alongside the general availability of Microsoft Discovery, we are introducing the Microsoft Discovery app available in preview today. The Microsoft Discovery app is a localized experience that gives researchers, students, academic labs, and scientific teams a simpler way to begin using Microsoft Discovery capabilities without starting with a full enterprise deployment. It is available for download on the Microsoft Discovery GitHub and users can get started with a GitHub Copilot account . This preview extends Microsoft Discovery to earlier stages of exploration, where research ideas begin as small-team projects, academic work, or individual investigation. The Microsoft Discovery app is designed to lower the barrier to hands-on exploration, with a practical entry point for literature exploration, hypothesis generation, scientific reasoning, and iterative experimentation. The app lets researchers explore Microsoft Discovery capabilities using their own working environment. As projects mature and complexity increases, researchers and teams can bring work developed locally into Microsoft Discovery platform to support more advanced R&D programs. Figure 4: The Microsoft Discovery app welcome screen . Applying Microsoft Discovery across R&D During preview, organizations helped shape the path to general availability by sharing feedback on how they were using Microsoft Discovery to explore advanced R&D workflows grounded in domain-specific data, established research methods, and expert review. Partners are contributing domain expertise and solution depth that can help organizations adapt Microsoft Discovery to the tools, data, and processes already central to their R&D work. Together, this work offers an
Originally published at azure.microsoft.com
