Accordingly, the two platforms AutoGen and Semantic Kernel will switch to maintenance mode, only accepting patches and bug fixes instead of continuing to develop new features.
Agent framework is designed as a single framework that allows businesses to build, deploy, and monitor a wide range of AI agents in the same system. The new framework closely integrates with Azure AI Foundry, supports pre-deployment local testing, connects API ( Application Programming Interface) via OpenAPI, manages multiple agents at the same time and reduces the need to switch between different platforms.
Agent framework's highlight is its focus on safety, security, and visibility. The system can detect sensitive data (PII), prevent prompt injection attacks, and provide a performance and actor behavior tracking tool based on OpenTelemetry opening standards.
According to Ms. Sarah Bird - Responsible AI Product Director at Microsoft, businesses need to consider three main factors when applying AI Agency: quality, security and management capability. Agent framework was born to comprehensively address all three aspects, aiming to deploy thousands of actors in the business environment safely and effectively.
With this step, Microsoft expects to strengthen its position in the AI Agency race, competing directly with other platforms such as LangChain, crewAI or LlamaIndex. At the same time, it creates an advantage thanks to its deep integration into the Azure ecosystem and Microsoft 365.
