- Microsoft's strategy focuses on autonomous and collaborative AI agents in an open "web agentic."
- The open MCP protocol and new standards for interoperability and security between agents are being promoted.
- GitHub Copilot, Azure Foundry, and Windows AI Foundry are evolving to offer environments for building, orchestrating, and managing AI agents in the cloud and on-premises.
- Microsoft incorporates identity, memory, personalization, and governance mechanisms, as well as tools to accelerate scientific and business innovation.
Artificial intelligence has taken a key leap forward at the conference Microsoft Build 2025, where the company presented its vision of the New “web agentic”: an open network of autonomous AI agents, capable of collaborating, making decisions, and acting without direct supervisionThis open standards-based approach promises to transform how people and businesses interact with technology.
Microsoft is committed to an open, collaborative, and secure approach, integrating solutions into its ecosystem to facilitate the adoption of intelligent agents. The company doesn't focus solely on tools or the cloud: Its proposal ranges from the Windows desktop to the business world, through scientific projects and community development itself. I'll tell you.
The 'Agentic Web': From Reactive Assistants to Autonomous and Connected Agents
Microsoft's vision for web agents goes far beyond traditional virtual assistants. These agents not only respond to specific requests, but can also anticipate and coordinate with other agents., and even tackle complex tasks in different environments and platforms. The key is to provide them with greater autonomy and a mechanism that allows them to exchange information and collaborate efficiently.
To achieve this, Microsoft is betting on the Model Context Protocol (MCP), an open standard created by Anthropic that facilitates interoperability between agents from different companies and AI models. The goal is to avoid technological silos and create an ecosystem where any agent can integrate and collaborate, regardless of their origin or platform.
The company has also unveiled the project NLWeb, which will allow websites to incorporate conversational interfaces accessible to different agents, similar to how HTML enabled the universalization of the web in the 1990s.
Microsoft platforms for creating AI agents with augmented intelligence
Microsoft has revamped its portfolio of solutions to facilitate the full lifecycle of AI agents, from ideation to management and customization. Key announcements include:
- GitHub Copilot It's evolving to become more than just a code assistant, acting as a "digital" member of the development team, capable of performing tasks, collaborating with other stakeholders, and participating in multiple phases of software development. Its chat is also being launched in Visual Studio Code to foster open innovation.
- Azure AI Foundry It integrates its own and third-party models (such as xAI's Grok models), as well as new tools for evaluating and selecting models, supporting multi-agent workflows, integration with SharePoint, and open protocols such as MCP and A2A.
- Windows AI Foundry provides a unified platform for developing and running local AI, making it easy to optimize models and agents on devices with varied hardware architectures.
- Copilot Studio and Copilot Tuning They allow agents to be created and adjusted using internal company data and processes, even with little or no programming knowledge, facilitating customization and adjustment to specific tasks.
The approach focuses on supporting both local and cloud development, enabling the deployment of agents in Windows, Mac, and cloud environments, as well as facilitating their integration into various business workflows.
Digital identity and memory: towards smarter and more reliable agents
One of the most talked about new features is the introduction of Microsoft Enter Agent ID, a system whereby each agent generated through Copilot Studio or Azure Foundry receives a unique digital identity. This facilitates the management of permissions, access and governance in complex business environments., by equating agents to “digital employees” with defined rights and limitations.
Another relevant point is the work in improve the memory of agentsMicrosoft has introduced structured memory technologies based on Retrieval-Augmented Generation (RAG), which allow agents to recall previous interactions and act accordingly. This paves the way for more personalized, efficient, and human-like experiences.
Additionally, security and compliance controls integrated with Microsoft Purview were introduced, as well as data loss prevention and observability mechanisms to monitor the actions and performance of each agent.
Practical applications: scientific discovery, automation, and new work models
Microsoft has demonstrated the potential of AI agents beyond the business or development arena. In the field of scientific research, The Microsoft Discovery Platform has already accelerated projects such as the discovery of new materials for data center cooling, reducing years of work to just days of automated analysis..
In the field of process automationMulti-agent orchestration offers a solution for tackling complex tasks through delegation and collaboration between specialized systems. This can transform how teams work and how data and information flows are managed in organizations across any industry.
Finally, the integration of open standards and the commitment to interoperability mark the way towards an ecosystem where Artificial intelligence not only responds to queries, but anticipates needs, proposes solutions, and collaborates securely and flexibly. with other agents inside and outside the organization.
The commitment to web agents represents a step forward in the evolution of artificial intelligence: Microsoft proposes a future where AI agents are protagonists, collaborate with each other transparently and help accelerate innovation. in different areas. All of this, under an open, safe approach designed for collective benefit, both in professional and personal settings.
Frequently asked questions about Microsoft Web Agent
- What is Web Agentic, introduced by Microsoft? It is an open network of artificial intelligence agents capable of collaborating, making decisions, and acting autonomously in different technological environments.
- What role does Azure AI Foundry play in web agents? It allows you to integrate and manage multiple AI models and agents, facilitating their orchestration both in the cloud and on-premises environments.
- How are Microsoft AI agents identified? Through Entra Agent ID, a system that assigns unique digital identities to each agent, improving security and governance in companies.
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.