We’re proud to share that Microsoft has as soon as once more been named a Chief within the 2025 Gartner® Magic Quadrant™ for Knowledge Science and Machine Studying (DSML) Platforms.
We’re proud to share that Microsoft has as soon as once more been named a Chief within the 2025 Gartner® Magic Quadrant™ for Knowledge Science and Machine Studying (DSML) Platforms. We imagine this recognition displays our continued dedication to offering organizations with a complete toolchain for constructing and deploying machine studying fashions and AI purposes, reworking how companies function. Azure Machine Studying is a part of a broad, interoperable ecosystem throughout Microsoft Cloth, Microsoft Purview, and inside Azure AI Foundry.
Gartner defines a knowledge science and machine studying platform as an built-in set of code-based libraries and low-code tooling. These platforms assist the unbiased use and collaboration amongst knowledge scientists and their enterprise and IT counterparts, with automation and AI help by way of all levels of the information science life cycle, together with enterprise understanding, knowledge entry and preparation, mannequin creation, and sharing of insights. Additionally they assist engineering workflows, together with the creation of knowledge, characteristic, deployment, and testing pipelines. The platforms are offered through desktop shopper or browser with supporting compute situations or as a completely managed cloud providing.
Main the way in which in 2025
With Microsoft, we’re turning our media experience right into a aggressive benefit—and harnessing knowledge to construct manufacturers and drive enterprise progress.
—Callum Anderson, World Director for DevOps and SRE at Dentsu.
At Microsoft, we envision a unified expertise the place knowledge scientists, AI engineers, builders, IT operations professionals, and enterprise customers come collectively to create purposes and handle the complete AI lifecycle throughout personas and initiatives. To that finish, in November 2024, we introduced the supply of Azure AI Foundry—a platform that permits builders to design, customise, and handle AI purposes. Azure Machine Studying is a trusted workbench that exists on prime of Azure AI Foundry and powers the underlying device chain expertise, with capabilities for mannequin customization, together with fine-tuning and RAG.
Advancing AI with Azure Machine Studying and clever brokers
As a part of Azure AI Foundry, the Foundry Agent Service empowers developer groups to orchestrate AI brokers that automate complicated, cross-functional workflows. Whether or not constructing options for software program engineering, enterprise course of automation, buyer assist, or knowledge evaluation, Foundry Agent Service offers a strong, safe, and interoperable basis to operationalize AI brokers in manufacturing environments.
- With assist for multi-agent orchestration, builders can design agent techniques that coordinate throughout duties, share state, get better from failures, and evolve flexibly as necessities change. These brokers will be grounded in enterprise data utilizing Microsoft Cloth, Bing, and SharePoint, whereas interacting with each proprietary and third-party instruments because of open requirements like MCP (Mannequin Context Protocol) and A2A (Agent2Agent).
- Builders can begin constructing regionally utilizing open-source frameworks like Semantic Kernel and AutoGen, and we’re on a transparent path towards delivering a unified SDK throughout the 2 frameworks and Azure AI Foundry that means that you can transfer from native experimentation to manufacturing in cloud with out rewriting any code. This ensures constant developer expertise—from preliminary prototyping to managed orchestration with observability and enterprise-grade management.
Collectively, Azure Machine Studying and Foundry Agent Service allow a future the place AI techniques are designed for enterprise use with scalability and safety in thoughts.
Leveraging AI fashions with Azure AI Foundry
Azure AI Foundry affords builders an progressive technique of deploying and managing its over 11,000 AI fashions with instruments just like the Mannequin Router, Mannequin Leaderboard, and Mannequin Benchmarks.
- The Mannequin Leaderboard simplifies the comparability of mannequin efficiency throughout real-world duties, offering clear benchmark scores, task-specific rankings, and reside updates, enabling customers to pick the excessive accuracy, quick throughput, or aggressive price-performance ratio effectively.
- Mannequin Benchmarks in Azure AI Foundry provide a streamlined technique to evaluate mannequin efficiency utilizing standardized datasets, whereas additionally permitting clients to judge fashions on their very own knowledge to determine one of the best match for his or her particular eventualities.
- Complementing this, the Mannequin Router—accessible now for Azure OpenAI fashions—dynamically routes queries to probably the most appropriate giant language mannequin (LLM) by assessing elements equivalent to question complexity, price, and efficiency, guaranteeing high-quality outcomes whereas minimizing compute bills.
These capabilities empower companies to deploy versatile and adaptive AI techniques with enterprise-grade efficiency, safety, and governance. With built-in innovation from Microsoft and its ecosystem, customers achieve entry to future-ready options that improve effectivity and scalability, guaranteeing they keep forward within the quickly evolving AI panorama.
Optimizing AI efficiency with fine-tuning in Azure AI Foundry
Effective-tuning is a vital device for organizations aiming to customise pre-trained AI fashions for particular duties, enhancing their efficiency, accuracy, and adaptableness, all whereas decreasing operational prices. Effective-tuning in Azure AI Foundry is powered by the underlying Azure Machine Studying device chain.
- With improvements equivalent to Reinforcement Effective-Tuning (RFT) utilizing the o4-mini mannequin, Azure AI Foundry allows builders to enhance reasoning, context-aware responses, and dynamic decision-making by way of reinforcement alerts. This adaptability is especially suited to purposes requiring ongoing studying, making it a perfect technique for evolving enterprise logic and guaranteeing fashions keep related in dynamic environments.
- Azure AI Foundry additional simplifies fine-tuning with options equivalent to World Coaching and the Developer Tier. World Coaching lowers prices by permitting mannequin customization throughout a number of Azure areas, giving builders flexibility and scalability whereas adhering to strict privateness insurance policies. The Developer Tier affords an inexpensive technique to consider fine-tuned fashions, enabling simultaneous testing throughout deployments and empowering customers to decide on one of the best candidate for manufacturing with precision and effectivity.
Collectively, these capabilities allow builders and enterprises to unlock the complete potential of their AI techniques, driving innovation and effectivity within the quickly evolving digital panorama.
Enabling organizations to deploy AI options
From healthcare and finance to manufacturing and retail, clients are utilizing Azure Machine Studying to unravel complicated issues, optimize operations, and unlock new enterprise fashions. Whether or not it’s deploying basis fashions, orchestrating AI brokers, or scaling real-time inference, Microsoft helps organizations flip knowledge into influence.
Start your journey with Azure Machine Studying
The migration to Azure is only the start. We’ve laid the inspiration to discover alternatives we might solely think about earlier than.
—Steve Fortune, Chief Digital and Expertise Officer at CSX.
Machine studying is revolutionizing the operational and aggressive panorama for companies within the digital age. It affords alternatives to optimize enterprise processes, enhance buyer experiences, and drive innovation. Azure Machine Studying serves as a strong and versatile platform for machine studying and knowledge science, enabling organizations to implement AI options responsibly and successfully.
Gartner, Magic Quadrant for Knowledge Science and Machine Studying Platforms, By Afraz Jaffri, Maryam Hassanlou, Tong Zhang, Deepak Seth, Yogesh Bhatt, 28 Might 2025.
GARTNER is a registered trademark and repair mark of Gartner, Inc. and/or its associates within the U.S. and internationally, Magic Quadrant is a registered trademark of Gartner, Inc. and/or its associates and is used herein with permission. All rights reserved.
This graphic was printed by Gartner, Inc. as half of a bigger analysis doc and needs to be evaluated within the context of the complete doc. The Gartner doc is obtainable upon request from [https://www.gartner.com/en/documents/6533902].
Gartner doesn’t endorse any vendor, services or products depicted in its analysis publications, and doesn’t advise expertise customers to pick solely these distributors with the very best rankings or different designation. Gartner analysis publications encompass the opinions of Gartner’s analysis group and shouldn’t be construed as statements of truth. Gartner disclaims all warranties, expressed or implied, with respect to this analysis, together with any warranties of merchantability or health for a selected function.