.svg){kind=spacer}
Details
- At Build 2025 in May, Microsoft introduced a wide-ranging AI agent ecosystem, unveiling GitHub Copilot agent mode, updates to Azure AI Foundry, and the new Windows AI Foundry.
- The new platforms support agent orchestration with the Azure AI Foundry Agent Service, offer model performance comparison via a Model Leaderboard, and enable secure agent identity management through Microsoft Entra Agent ID.
- Integration of Grok-3 AI models from xAI and the NLWeb protocol aim to standardize how agents interact with the web, mirroring the foundational impact HTML had on the early internet.
- Microsoft's open agentic web vision allows AI agents to carry out tasks across different platforms while upholding organizational security and compliance using Purview controls.
- Early enterprise adopters include Fujitsu and NTT Data, leveraging Azure AI Foundry for sales optimization, and Stanford Health Care, which is applying the technology to automate medical workflows.
Impact
This announcement underscores Microsoft’s ambitions to become the leading infrastructure provider for enterprise AI, directly challenging offerings from AWS Bedrock and Google Vertex AI. The adoption of the NLWeb protocol could drive broader interoperability for agent-based applications, while local and cloud model support answers shifting needs for edge computing. Microsoft's strategy sets the stage for widespread enterprise deployment of AI agents through 2025 and beyond.
