- Microsoft Discovery AI transforms R&D with intelligent agents and open platforms.
- Azure AI Foundry, NLWeb, and GitHub Copilot are evolving, integrating new models and protocols.
- Security, memory, and openness are key to Microsoft's new strategy.
Microsoft Discovery AI has become the epicenter of the digital transformation for scientific institutions, companies, and developers looking to boost their research and development projects. During its recent global conferences and events, the firm has presented a vision that goes beyond simple virtual assistants, betting on open networks of intelligent agents capable of collaborating, learning and executing complex tasks both in local and cloud environments.
The advancement of these technologies represents an opportunity for accelerate the creation of new products, facilitate decision-making, or shorten the timescale for scientific discoveries. The focus is on developing open, interoperable, and secure platforms that allow companies of all sizes and research laboratories to access the latest innovations in agentic artificial intelligence.
Open ecosystem: from Azure AI Foundry to NLWeb
One of the pillars of this evolution is Azure AI Foundry, an environment in which developers can create, train, and deploy intelligent agents in a unified manner. New integrations of language models stand out, such as Grok 3 y Grok 3 mini (developed by xAI, Elon Musk's company), in addition to compatibility with third-party models such as Mistral and Llama, and a strengthened collaboration with hugging face which expands access to over 11.000 models directly in the Microsoft cloud. Although you can also explore how turn your PC into a local AI hub.
The platform introduces tools such as Model Leaderboard to compare models based on their performance and Model Router to automatically select the most appropriate model for each need or query. This flexibility makes Azure AI Foundry a infrastructure capable of supporting more than 1.900 models, enhancing the diversity and depth of applications in R&D, healthcare, education and other key sectors.
Complementing this approach, Microsoft has launched NLWeb, a new standard designed to transform any website into a conversational space accessible by AI agents. Its goal is to facilitate interaction between users, agents, and web content, playing a similar role to that played by HTML in the era of the visual web. NLWeb, researchers can use customized intelligent assistants to explore databases or extract specialized information in real time.
Windows AI Foundry and Local Agents
In addition to cloud services, Windows AI Foundry emerges as a solution for integrating and running artificial intelligence models directly on local devices. Developers and scientists can take advantage APIs ready for vision and language tasks, work with open source models, and tune and deploy models both on-premises and in cloud environments.
In parallel, the official release of the Windows Subsystem for Linux (WSL) as an open source project, encouraging collaboration with the community and the creation of new features within the Microsoft development ecosystem.
Open protocols and collaboration between agents
The architecture proposed by Microsoft is based on interoperability. Model Context Protocol (MCP), called by the company itself as the “USB-C of AI”, allows intelligent assistants and agents from different platforms to communicate and collaborate with each other. Integrates logging mechanisms and MCP servers, ensuring that interactions between systems are context-preserved and secure.
This open approach extends to platforms such as GitHub, Dynamics 365, Copilot Studio and Windows 11Thanks to MCP and the new generation of agent services, different applications can exchange information, delegate tasks, and maintain a contextualized experience tailored to both individual users and large organizations.
Practical applications of Discovery AI in science, health, and education
Microsoft Discovery AI has been presented as a key platform for accelerate R&D processes in hospitals, universities and research centersAt recent events, the company has showcased examples such as the accelerated development of new drugs, the rapid identification of sustainable materials (such as PFAS-free refrigerants for data centers), and the optimization of clinical operations through multi-agent orchestrators.
In education, institutions such as the World Bank have leveraged personalized artificial intelligence to improve learning in schools across different countries, adapting tools to the needs of neurodivergent students and facilitating the creation of large-scale educational resources.
I am a technology enthusiast who has turned his "geek" interests into a profession. I have spent more than 10 years of my life using cutting-edge technology and tinkering with all kinds of programs out of pure curiosity. Now I have specialized in computer technology and video games. This is because for more than 5 years I have been writing for various websites on technology and video games, creating articles that seek to give you the information you need in a language that is understandable to everyone.
If you have any questions, my knowledge ranges from everything related to the Windows operating system as well as Android for mobile phones. And my commitment is to you, I am always willing to spend a few minutes and help you resolve any questions you may have in this internet world.