Navigating Microsoft's AI Landscape: From Microsoft.Extensions.AI to Semantic Kernel

It seems increasingly clear that Semantic Kernel is where Microsoft is investing for building modern, robust AI applications. In fact, Microsoft has stated as much: “Semantic Kernel (is) central to Microsoft’s enterprise AI story”.

In typical Microsoft fashion, they offer several seemingly competing options for developers to choose from, such as AutoGen, Microsoft.Extensions.AI, and probably a few others I’m forgetting right now. This variety doesn’t make it any easier for .NET developers to figure out where to get started and where to focus their upskilling efforts.

I plan to focus on the foundational abstractions in Microsoft.Extensions.AI and “upgrade” to Semantic Kernel when needed. The previously mentioned post seems to indicate that this is Microsoft’s direction too.

The road may be a bit bumpy as Microsoft starts to consolidate their developer frameworks, so hold on! I’m looking forward to the journey!

P.S. There are lots of exciting, enterprise-grade innovations happening on the “low code” Copilot Studio side of the house too, such as Copilot Tuning, multi-agent orchestration, Entra Agent ID, and M365 Agents Toolkit! Exciting times!