What you'll learn:
- Design a Microsoft Foundry resource and project structure for enterprise AI agents.
- Select and deploy Foundry models using identity, quota, regional, latency, and cost constraints.
- Build, version, publish, and invoke governed agents through Foundry Agent Service.
- Ground agents with Foundry IQ and Azure AI Search using cited, permission-aware retrieval.
- Connect agents to approved tools and enterprise systems through catalog, OpenAPI, and MCP patterns.
- Orchestrate specialist .NET agents with Microsoft Agent Framework workflows, fan-in/fan-out patterns, and A2A communication.
- Apply Azure Policy, guardrails, AI Gateway, identity, red teaming, and Control Plane governance.
- Evaluate, trace, monitor, and release agents with measurable quality, safety, latency, and cost gates.
Microsoft Foundry for Enterprise AI Agents is designed for experienced developers such as AI engineers, .NET developers, and ML engineers seeking to build enterprise-grade AI solutions on Azure. This course assumes familiarity with cloud development, AI concepts, and .NET, ensuring learners are prepared to engage with advanced topics effectively. The course centers on a realistic enterprise scenario: the Northstar Operations Assistant, a practical project that demonstrates how AI agents can investigate issues, retrieve enterprise evidence, coordinate tools, and support traceable decisions-skills directly applicable to enterprise AI deployments.
Build Governed Agents with Foundry Agent Service and .NET.
This course focuses on the architecture and operating model required when AI agents interact with enterprise data, identities, APIs, tools, and production controls. You will build and invoke agents with Foundry Agent Service and the Responses API. The course explains agent types, instructions, conversations, versioning, publishing, safe failure, and human approval. You will test both supported and unsupported requests so you can evaluate agent behavior rather than assume it.
Ground Enterprise Agents with Foundry IQ and RAG
Enterprise agents need evidence, not simply fluent output. You will explore Retrieval-Augmented Generation (RAG), the difference between classic RAG and agentic retrieval, and how Foundry IQ provides a managed knowledge layer for AI agents. You will work with knowledge sources, knowledge bases, Azure AI Search, permission-aware retrieval, and citations. You will also test unsupported questions and examine why a citation must support the actual claim rather than merely reference a related source. The objective is to build agents that can distinguish between evidence-backed answers and responses that should be refused or escalated.
Why This Course Is Different
Many AI agent courses focus on isolated features like agent creation or tool calling. This course links these capabilities to enterprise concerns like identity, authorization, evidence, and fleet governance, helping professionals build skills for responsible AI deployment and career growth.
The implementation path also gives .NET developers a first-class route through Microsoft's AI stack. You will see how C# and .NET fit with Microsoft Foundry and Microsoft Agent Framework rather than treating Python as the only development option. The goal is not to reproduce every feature available in the portal. The goal is to understand the decisions that make an AI agent architecture secure, reviewable, measurable, and maintainable.
Technologies and Concepts Covered
Microsoft Foundry,
Azure AI Foundry terminology,
Foundry Agent Service,
Foundry IQ,
Microsoft Agent Framework,
.NET 10,
C#,
Responses API,
Azure AI Search,
Retrieval-Augmented Generation (RAG),
Model Context Protocol (MCP),
OpenAPI,
Agent-to-Agent (A2A),
Microsoft Entra ID,
RBAC,
Managed Identity,
Application Insights,
OpenTelemetry,
LLMOps,
Governance.
Who This Course Is For
This course is designed for AI engineers, .NET developers, cloud developers, ML engineers, architects, platform engineers, and technical leads who need to build or review enterprise AI agent solutions on Azure.
It is particularly relevant if you need to answer production questions such as how Microsoft Foundry resources should be organized, which identity an agent should use, how enterprise knowledge should be grounded and cited, which tools an agent should be permitted to call, when multiple specialist agents are justified, how agent behavior should be evaluated before release, how you can trace what an agent actually did, and which safety, policy, cost, and governance controls belong around the system.
By the end of the course, you will have a reusable mental model for designing enterprise AI agents that are not only capable, but also explainable, traceable, measurable, secure, and governable.
If you want to understand how Microsoft Foundry, .NET, Foundry IQ, MCP, Microsoft Agent Framework, LLMOps, observability, and enterprise governance fit together in one production-oriented agent architecture, this course takes you through that complete lifecycle.
