Agent Orchestration with Microsoft Agent Framework

In this tutorial, we go beyond basic workflows in the Microsoft Agent Framework and explore multi-agent orchestration. You’ll learn how to route between multiple agents using conditional logic, switch statements, and predicate factories, enabling intelligent multi-step AI workflows.

We start with a simple two-agent setup, then scale it to handle multiple agents—including tourist recommendations, restaurant retrieval from RAG, and even a hockey agent (because… Canada). By the end, you’ll understand how to control agent routing based on weather, country, and more.

This video is part of my ongoing series on the Microsoft Agent Framework. More tutorials coming soon — subscribe so you don’t miss out.

What you’ll learn:

How to orchestrate multiple agents
Building conditional edges in workflows
Implementing switch-style routing logic
Integrating RAG-based tools into agents
Expanding city info models for decision-making
Practical examples: weather-based routing, country-based switches, and more.

🔗 Code Repo:
https://github.com/kirkmcpherson/micr...

🔗 Previous Video:
   • Agent Workflows with Microsoft Agent Frame...  

00:00 — Intro: Scaling Beyond Two Agents
00:30 — Overview of Multi-Agent Orchestration
00:56 — Recap: Workflows & City Info Example
01:48 — Adding More Agents (RAG + Restaurant Recommendations)
03:32 — Implementing Conditional Logic (Sunny vs Not Sunny)
06:33 — Building Conditional Edges in the Workflow
11:03 — Introducing Switch-Style Routing
12:07 — Creating the Hockey Agent + Predicate Factory
14:31 — Switch Case Workflow: Canada, USA, and Default
16:01 — Testing: Montreal, New York, Paris
18:56 — Final Thoughts & What’s Coming Next

#MicrosoftAgentFramework #AIAgents #AgentOrchestration #AIWorkflows #MultiAgentSystems #RAG #AzureAI #LLMApps #AIDevelopment #ArtificialIntelligence #TechTutorial #CodingTutorial #SoftwareEngineering #PythonDevelopers #AIWorkflow