Session Recap: The Evolution of .NET AI — From Raw Requests to Autonomous Agents

On October 8th I was back at the Cyber RGV monthly meetup at the McAllen Public Library for my second talk, this time on something I've been deeply into lately: building AI-powered applications with .NET. The talk walks through five distinct phases of how AI integration in the .NET ecosystem has evolved, starting from the ugly-but-honest approach of hand-rolling raw HTTP requests to every AI API, all the way to building stateful, multi-agent workflows with the newly released Microsoft Agent Framework (MAF).

The through-line is how each layer (Semantic Kernel, Extensions.AI, AutoGen.NET, and finally MAF) builds on the last to reduce friction and raise the ceiling on what you can actually build. By the end we're looking at a fully stateful customer service agent that handles tickets, tracks orders, and generates emails, built in essentially three lines of code, with conversations that can be serialized, saved, and picked back up months later with full context intact. It's the kind of thing that makes you realize .NET isn't just keeping pace with the AI wave — it's got a genuinely mature ecosystem for riding it.

The full session is on YouTube — if you're a .NET developer trying to figure out where to start with AI, or you've started but feel like you're fighting the tooling more than building things, I think this one's worth your time.

Watch: The Evolution of .NET AI — From Raw Requests to Autonomous Agents

And if you're in the RGV, I'd love to see you at the first official RGV Devs Dot Net session on April 14 — we're kicking things off with a deep dive into building an AI brand engine with .NET and MAF:

RGV Devs Kickoff — Building an AI Brand Engine with .NET and MAF