Announcing Microsoft Discovery general availability and Microsoft Discovery app preview

Breakthroughs successful subject and engineering seldom travel from a azygous insight. They look done cycles of hypothesis, experimentation, refinement, and reappraisal crossed teams, tools, and data.

Today astatine Microsoft Build, we are announcing that Microsoft Discovery is present mostly available for each organizations, providing a broad level for gathering and governing agentic AI workflows crossed technological and engineering disciplines. We are besides introducing the Microsoft Discovery app successful preview, a section desktop acquisition that helps researchers, students, and technological teams statesman moving with Microsoft Discovery today.

Since introducing Microsoft Discovery successful backstage preview at Microsoft Build past year, we person worked intimately with organizations applying AI to analyzable probe and improvement (R&D) workflows. Their feedback helped reenforce wherever agentic AI needs to spell beyond idiosyncratic assistance, similar supporting the iterative loops, grounds preservation, and instrumentality coordination that specify technological work.

The astir challenging problems successful R&D necessitate much than conscionable a punctual interface oregon a azygous exemplary response. Scientific workflows require:

• Integration with organization cognition and domain expertise.
• Access to specialized modeling, simulation, and investigation tools.
• Connection to experimental grounds and validation data.
• Support for reappraisal processes that signifier probe decisions.

A materials idiosyncratic whitethorn request to measure performance, safety, and outgo alongside manufacturability and regulatory constraints. A semiconductor squad whitethorn request to research a larger plan abstraction without losing carnal fidelity oregon traceability. A beingness sciences researcher whitethorn request to link lit and experimental information with models and cohort-level grounds earlier deciding what to validate next.

Microsoft Discovery is designed to enactment wrong these existing R&D environments, not regenerate them. The level helps experts recognize the reasoning way down outputs and keeps quality judgement astatine the halfway of technological and engineering decisions. The wide availability of Microsoft Discovery marks a important milestone successful turning these requirements into a production-ready level for R&D environments with governance and transparency built in.

Figure 1: The Microsoft Discovery workspace invited experience.

How Microsoft Discovery supports R&D workflows astatine scale

Microsoft Discovery enables organizations to specify agentic workflows astir their ain R&D programs. Teams tin make and coordinate specialized agents, link those agents to organization cognition and outer technological information, and orchestrate enactment crossed modeling, simulation, analysis, and validation tools.

At the halfway of the level is the Microsoft Discovery Engine, which supports the halfway loop of technological enactment by helping teams determination from grounds to hypotheses, done execution and analysis, and into the adjacent iteration. This loop allows teams to determination beyond isolated investigation toward repeatable, evidence-driven exploration, wherever they tin comparison tradeoffs, question assumptions, and constrictive a hunt abstraction successful a mode that tin beryllium reviewed and repeated.

Figure 2: The Microsoft Discovery Engine task instauration and presumption overview.

As we continued merchandise development, we focused connected what it takes to bring agentic AI into accumulation R&D environments:

• Workflows request to stay reproducible.
• Outputs indispensable beryllium reviewable.
• Proprietary cognition indispensable beryllium connected and governed appropriately.
• Agentic systems request to acceptable into the operating exemplary of R&D organizations.

Those considerations, on with continued lawsuit feedback, helped signifier the wide availability merchandise and the level capabilities down it.

Figure 3: The Microsoft Discovery Engine output with assurance scoring and cited probe findings.

Expanding entree with the Microsoft Discovery app preview

An important extremity for Microsoft is to marque precocious AI and computing capabilities much accessible to the radical moving connected immoderate of today’s astir hard technological and engineering challenges. Alongside the wide availability of Microsoft Discovery, we are introducing the Microsoft Discovery app disposable successful preview today.

The Microsoft Discovery app is simply a localized acquisition that gives researchers, students, world labs, and technological teams a simpler mode to statesman utilizing Microsoft Discovery capabilities without starting with a afloat endeavor deployment. It is disposable for download connected the Microsoft Discovery GitHub and users tin get started with a GitHub Copilot account.

This preview extends Microsoft Discovery to earlier stages of exploration, wherever probe ideas statesman arsenic small-team projects, world work, oregon idiosyncratic investigation. The Microsoft Discovery app is designed to little the obstruction to hands-on exploration, with a applicable introduction constituent for lit exploration, proposal generation, technological reasoning, and iterative experimentation.

The app lets researchers research Microsoft Discovery capabilities utilizing their ain moving environment. As projects mature and complexity increases, researchers and teams tin bring enactment developed locally into Microsoft Discovery level to enactment much precocious R&D programs.

Figure 4: The Microsoft Discovery app invited screen.

Applying Microsoft Discovery crossed R&D

During preview, organizations helped signifier the way to wide availability by sharing feedback connected however they were utilizing Microsoft Discovery to research precocious R&D workflows grounded successful domain-specific data, established probe methods, and adept review.

Partners are contributing domain expertise and solution extent that tin assistance organizations accommodate Microsoft Discovery to the tools, data, and processes already cardinal to their R&D work. Together, this enactment offers an aboriginal presumption into how Microsoft Discovery is being utilized crossed domains and however a increasing ecosystem tin assistance marque analyzable R&D workflows much systematic, transparent, and repeatable.

Yale Engineering

A collaboration crossed Professor David Kwabi’s radical astatine Yale Engineering and researchers from Microsoft utilized the Discovery Engine to beforehand the frontier of agentic tiny molecule plan for grid-scale aqueous integrated redox travel batteries (ORFBs).

ORFBs are promising, starring candidates for sustainable, environmentally friendly, long-duration vigor storage, but challenging to optimize. Electrolytes indispensable equilibrium analyzable molecular properties similar redox potential, aqueous solubility, synthetic tractability, and electrochemical reversibility. The Discovery Engine, gathering connected our cognitive loop via in-situ optimization research, enables long-horizon technological reasoning portion ensuring spot successful the full process.

With these capabilities, the squad utilized the agentic loop to thrust in-silico exploration and convergence of candidates, construe experimental results, and suggest diagnostic experiments. Experts astatine Yale Engineering led each experimental characterization, verified results interpretation, and evaluated the applicable applicability of the designs. The probe is disposable here.

This enactment introduces a almighty caller model for advancing artillery subject with AI. By endowing an cause with the quality to crushed from and accommodate to experiments, we harvester the strengths of human-led experimentation with AI’s capableness to research immense chemic plan spaces – and we’re lone opening to spot what it tin do.

—David Kwabi, Associate Professor, Yale

Georgia Institute of Technology

Georgia Tech is exploring however an agentic AI strategy tin re-evaluate the prebiotic plausibility of histidine, a biochemically important amino acerb whose emergence nether plausible prebiotic conditions remains unclear contempt its ubiquity successful biology. Classical instrumentality learning and AI approaches person struggled successful this domain owed to the deficiency of standardized datasets and the inherently multimodal quality of the data.

The projected script requires a multi-agent AI strategy composed of specialized AI ‘scientists’ for chiseled information modalities, including wide spectrometry analysis, lit extraction, planetary ngo information retrieval, and chemic absorption pathway modeling.

These agents volition collaborate done a cardinal reasoning coordinator to integrate divers and heterogeneous datasets, aiming to determination from “absence-of-evidence” to a robust, evidence-based appraisal of histidine’s prebiotic viability. The model developed tin besides beryllium repurposed to analyse different contested biosignatures, gathering a scalable pipeline for origins-of-life inquiry.

Our collaboration with the Microsoft Discovery squad done the Georgia Tech AI for Research programme has been highly valuable, some scientifically and operationally. Working unneurotic connected agentic AI systems to probe questions astir the origins of beingness has fixed america aboriginal vulnerability to the authorities of the creation embodied successful the Discovery platform, portion besides enabling genuinely adjacent method collaboration. This hands-on concern has enabled meaningful bidirectional learning.

—Dr. Amirali Aghazadeh, Assistant Professor, School of Electrical and Computer Engineering, Georgia Tech

Pacific Northwest National Laboratory

Microsoft and Pacific Northwest National Laboratory (PNNL) are rewriting the rules of technological discovery, unleashing AI that doesn’t conscionable assistance researchers but orchestrates the full find travel from caller hypotheses to real-world experiments.

Powered by Microsoft Discovery, cutting-edge robotics and AI agents enactment similar a virtual probe team: imagining experiments, reasoning crossed mountains of technological data, designing brand-new molecules, and learning connected the alert from unrecorded laboratory results astatine PNNL.

In vigor storage, this collaboration is fast-tracking the hunt for next-generation integrated redox travel artillery materials—breakthroughs that could slash our reliance connected captious minerals similar vanadium portion providing cheaper, much scalable vigor retention technologies that marque our powerfulness grid tougher than ever.

In biosystems engineering, Microsoft Discovery is plugging straight into PNNL’s laboratory automation infrastructure to motorboat self-driving technological workflows that autonomously design, run, and fine-tune biologic experiments successful existent time.

Together, Microsoft and PNNL are pioneering a caller exemplary for science, wherever robotics and autonomous laboratories fuse with AI and unreality infrastructure into 1 intelligent, closed-loop find motor that dramatically reduces the timeline from ideas to breakthroughs and opens the doorway to a caller epoch of innovation successful energy, biology, and worldly synthesis.

—Robert Runkle, Physicist and Lead for Autonomous Discovery Strategy, Pacific Northwest National Laboratory

Ginkgo Bioworks

Ginkgo Bioworks and Microsoft are collaborating to bring agentic AI into biologic discovery. Specialized agents tin analyse biologic datasets, make hypotheses, and plan experiments to execute connected an autonomous lab. Soon, researchers volition beryllium capable to scope and program experiments successful Microsoft Discovery and tally them straight connected Ginkgo Cloud Lab—no in-house automation required.

Together, agentic AI and autonomous labs volition alteration each portion of the technological process. Iteration cycles volition get faster, experiments volition necessitate little manual hands-on time, and computational analyses volition go much systematic and exhaustive. By making some easier to use, Microsoft and Ginkgo purpose to bring greater speed, standard and reproducibility to pre-clinical research.

—Jason Kelly, CEO, Ginkgo Bioworks, Inc.

Causaly

Causaly provides agentic solutions that compound the world’s biomedical grounds with an organization’s proprietary cognition to present confident, traceable, cited decisions astatine each stage, from find done launch.

Drug find does not endure from a deficiency of data. It suffers from a deficiency of trustworthy interpretation. Microsoft Discovery brings technological computation implicit endeavor data, and Causaly brings the anterior knowledge, mechanistic reasoning, and provenance needed to crook those signals into decisions. Together, we tin assistance researchers determination from earthy information to evidence-backed judgement overmuch faster and with greater confidence.

—Yiannis Kiachopoulos, Co-Founder and CEO, Causaly

Cambridge Consultants

With Microsoft Discovery, Cambridge Consultants is helping show however AI agents, simulation, and carnal laboratory systems tin enactment unneurotic successful a closed-loop find process.

These autonomous, AI-powered cycles tin crook months of experimental enactment into days oregon hours. The effect is simply a much connected exemplary for R&D, 1 designed to accelerate campaigner generation, experimental planning, and real-world validation.

Microsoft Discovery has the imaginable to assistance researchers determination faster from promising ideas to real-world results. We spot this arsenic an important measurement toward much scalable, integrated, and intelligent R&D.

—Joe Corrigan, Chief Technology Officer, Cambridge Consultants

Wiley

At each signifier of the probe and improvement process, beingness sciences and pharmaceutical teams request accelerated entree to the astir current, credible grounds available. Wiley Research Agent: Life Sciences delivers a continuously updated scale of much than 1 cardinal authoritative, high-quality, and trusted articles with hybrid hunt capabilities to enactment precocious technological reasoning.

The cause searches, retrieves, and synthesizes applicable findings into a coherent, evidence-based effect to queries. It tin run arsenic a stand-alone probe service, oregon successful orchestration with different Microsoft Discovery agents, fitting people into the broader technological reasoning workflows that Discovery enables. The Wiley Life Sciences Research Agent volition beryllium the archetypal of respective Wiley agents offered commercially connected the Microsoft Discovery level implicit time.

Scientific find depends connected connecting trusted grounds with progressively almighty AI systems. By bringing Wiley’s authoritative beingness sciences probe into Microsoft Discovery, we tin assistance beingness sciences and pharmaceutical teams accelerate proposal generation, experimentation, and results mentation crossed a continuous technological reasoning loop.

—Josh Jarrett, Senior Vice President and General Manager of Applied Research Intelligence astatine Wiley

BHP

BHP, the largest mining institution successful the world, is utilizing Microsoft Discovery to accelerate find of precocious copper leaching solutions—in a substance of months alternatively of years.

As copper request grows and caller deposits go harder to find and much costly to develop, improving betterment from existing ores is simply a captious lever to assistance conscionable aboriginal proviso needs. This concern has fixed our method experts the tools they request to constrictive an astir infinite tract of possibilities down to a tiny fig of options that could 1 time beryllium deployed successful our planetary copper operations. We are investigating against the realities of our ore bodies and operating constraints, truthful we are solving for what tin really enactment successful practice. This shows however exertion and quality expertise tin beryllium applied unneurotic to lick complex, real-world challenges.

—Jessica Farrell, Vice President Innovation, BHP

Syensqo

Syensqo is simply a planetary subject institution processing groundbreaking solutions that heighten the mode we live, work, travel, and play. The institution is presently leveraging Microsoft Discovery to standard agentic AI that accelerates discovery, improves decision-making, and unlocks measurable concern impact, peculiarly successful the improvement of next-generation vigor transportation fluids for semiconductor manufacturing.

We are present entering a caller signifier of our concern with Microsoft, focused connected scaling AI agents crossed research, income and selling to thrust near-term growth. By connecting lawsuit request to technological improvement and back-to-market execution, agentic AI is enabling faster cycles, sharper prioritization, and tangible interaction connected gross maturation and concern performance.

—Mike Radossich, CEO of Syensqo

GSK

GSK, the planetary biopharma, is moving to accelerate the discovery, development, and transportation of medicines and vaccines to patients.

Working with partners similar Microsoft Discovery, we spot the accidental to rapidly iterate connected campaigner molecules, perchance accelerating decision-making via accelerated information procreation and analysis.

—Christopher Austin, Senior Vice President, R&D Technologies, GSK

Figure 5: Microsoft Discovery has an expanding ecosystem of partners offering integrated tools and specialized expertise.

Microsoft Discovery is mostly available. The Microsoft Discovery app is disposable successful preview. Preview features and capabilities are taxable to change.