Microsoft Agent Framework

• Key points are in the visual diagram above.

Enterprises on Azure need agents with IAM, compliance, and .NET-first tooling. MAF unifies Semantic Kernel's plugin model with multi-agent patterns from AutoGen.

MAF is the corporate HR system for agents — standardized job descriptions (plugins), reporting lines (multi-agent), and payroll (Azure billing/monitoring).

Scenario

Enterprise copilot: SK agent with plugins for SharePoint search, Outlook send, and Teams notify — hosted on Azure AI Agent Service with Entra ID auth.

Solution

In Microsoft Agent Framework, apply Microsoft Agent Framework to this scenario: Enterprise copilot: SK agent with plugins for SharePoint search, Outlook send, and Teams notify — hosted on Azure AI Agent Service with Entra ID auth. Identify the inputs, run the technique, validate the output, and note one thing you would monitor in production.

Do this before moving to the next module — reading alone is not enough.

Open the Code Walkthrough below and run it locally. Change one parameter related to Microsoft Agent Framework (e.g. model, temperature, top_k, or tool name), observe the difference in output, and write 2–3 sentences explaining what changed.

Highlighted lines show where Microsoft Agent Framework happens in the code.

1from semantic_kernel.agents import ChatCompletionAgent  # import dependencies2from semantic_kernel import Kernel  # import dependencies3
4kernel = Kernel()5kernel.add_plugin(SharePointPlugin(), "sharepoint")6agent = ChatCompletionAgent(kernel=kernel, name="CorpAssistant",  # key line for Microsoft Agent Framework7    instructions="Search SharePoint and summarize for the user.")

• Treating Microsoft Agent Framework as a black box without evaluation
• Ignoring cost and latency in production
• Skipping error handling for microsoft agent framework

Quick recap — the most important points from this module.