Concepts
Concept 1 · 9 min read
What is Microsoft Foundry and why it was built
What Microsoft Foundry is, the words you need (resource, project, model, agent, tool, deployment), how the parts fit together, and what it costs to start.
Concept 2 · 11 min read
Where you work in Microsoft Foundry
How to work in Foundry: the portal, the SDK, the Azure Developer CLI, and the VS Code toolkit, with a first-hour setup from project to agent.
Concept 3 · 10 min read
The features of Microsoft Foundry
How Foundry's features fit together: agents, models, tools, tracing, evaluations, governance, and publishing, with preview features marked.
Concept 4 · 9 min read
Keys, costs, and deployment choices
Choose a deployment type, understand quota and token limits, pick between API keys and Microsoft Entra ID, and keep costs under control.
Concept 5 · 8 min read
Prompts and model settings in the playground
Use the Foundry playgrounds to test prompts, write a system message, ask for structured output, and change model settings with purpose.
Concept 6 · 9 min read
Knowledge with files and Azure AI Search
Give a Foundry agent answers from your own documents, using file search for a small set of files and Azure AI Search for a larger, indexed collection.
Concept 7 · 8 min read
Build a prompt agent with tools
Create a prompt agent from a model and instructions, add the tools it needs, control when it calls them, and save each milestone as a version.
Concept 8 · 9 min read
Measure and improve quality
Run a basic evaluation with clear evaluators, use Prompt Optimizer to improve instructions, and keep a baseline so every change is measured.
Concept 9 · 9 min read
Foundry architecture and rollout
How the Foundry resource, projects, and connected Azure services fit together, and how to plan a rollout that holds up under review.
Concept 10 · 9 min read
Models, deployments, and fine-tuning
Choose and govern models, manage model versions and retirements, and decide when fine-tuning is worth the effort.
Concept 11 · 11 min read
Agents, identity, workflows, and routines
Build production agents with their own identity, orchestrate several agents as workflows, and run agents on schedules or events.
Concept 12 · 10 min read
Tools, toolboxes, and knowledge
Build one governed toolbox endpoint for your agents, authenticate tools per user, and ground answers in your own knowledge.
Concept 13 · 9 min read
Observability and tracing
Trace every agent run with OpenTelemetry, monitor agents and deployments in production, and protect sensitive content in traces.
Concept 14 · 10 min read
Evaluation and red teaming
Build an evaluation workflow with the right evaluators, run it in your release pipeline, and red-team agents before users do.
Concept 15 · 11 min read
Guardrails, security, and network isolation
Apply guardrails at every intervention point, isolate networking, and control encryption keys for production agents.
Concept 16 · 10 min read
Governance, resilience, optimization, and publishing
Govern agents and tools at scale, design for resilience, optimize agent quality, and publish agents to Microsoft Teams and Copilot.
For professionals: get certified
Prove what you learned in this course with an official Microsoft certification. Prices depend on the country where you take the exam.
Associate · Exam AI-103
Microsoft Certified: Azure AI Apps and Agents Developer Associate
For developers who build AI apps and agents with Python on Microsoft Foundry.
- United States
- $165 USD
- India
- ₹4,865 INR
Price before tax. Microsoft sets the price by the country where you take the exam. Confirm the final amount at checkout.
View certification on Microsoft Learn ↗Fundamentals · Exam AI-901
Microsoft Azure AI Fundamentals
Start hereA starting point for anyone new to AI on Azure.
- United States
- $99 USD
- India
- ₹3,691 INR
Price before tax. Microsoft sets the price by the country where you take the exam. Confirm the final amount at checkout.
View certification on Microsoft Learn ↗
Sources
- What is Microsoft Foundry? (Microsoft Learn)
- Microsoft Foundry architecture (Microsoft Learn)
- Create a project (Microsoft Learn)
- Quickstart: Create a prompt agent (Microsoft Learn)
- Microsoft Foundry Playgrounds (Microsoft Learn)
- Agent development lifecycle in Microsoft Foundry (Microsoft Learn)
- Deploy Microsoft Foundry Models in the Foundry portal (Microsoft Learn)
- Deployment overview for Microsoft Foundry Models (Microsoft Learn)
- Role-based access control for Microsoft Foundry (Microsoft Learn)
- Run evaluations from the Microsoft Foundry portal (Microsoft Learn)
- Agent tracing overview (Microsoft Learn)
- Set up tracing for AI agents in Microsoft Foundry (Microsoft Learn)
- Review agent interactions with Trace Replay (Microsoft Learn)
- Use the Microsoft Foundry Skill in coding agents (Microsoft Learn)
- What is Foundry IQ? (Microsoft Learn)
- Microsoft Foundry: Scale innovation on a modular, interoperable, and secure agent stack (Azure blog, Nov 18, 2025)
- Microsoft Foundry pricing (Azure), including the FAQ on the free-to-explore platform
- Create your Azure free account or pay as you go (Azure)
- Deployment types in Microsoft Foundry Models (Microsoft Learn)
- Quotas and limits for Microsoft Foundry Models (Microsoft Learn)
- Enforce token limits for models (Microsoft Learn)
- Create and manage budgets in Azure Cost Management (Microsoft Learn)
- System message design
- Structured outputs
- File search tool
- Azure AI Search tool
- Tool best practices
- Evaluation in Microsoft Foundry
- Optimize agent prompts with Prompt Optimizer
- Optimize a prompt agent
- Foundry capability reference (Microsoft Learn)
- Plan rollout (Microsoft Learn)
- Foundry Models overview (Microsoft Learn)
- Model retirements (Microsoft Learn)
- Fine-tuning in Microsoft Foundry (Microsoft Learn)
- Agent identity concepts (Microsoft Learn)
- Workflows in Microsoft Foundry (Microsoft Learn)
- Routines in Foundry Agent Service (Microsoft Learn)
- Create and manage a toolbox in Foundry (Microsoft Learn)
- Tool authentication in Foundry toolboxes (Microsoft Learn)
- Govern MCP tools by using an AI gateway (Microsoft Learn)
- AI Red Teaming Agent (Microsoft Learn)
- Guardrails and controls overview (Microsoft Learn)
- Intervention points for guardrails (Microsoft Learn)
- Configure network isolation for Microsoft Foundry (Microsoft Learn)
- Customer-managed key encryption in Microsoft Foundry (Microsoft Learn)
- Foundry Control Plane overview (Microsoft Learn)
- High availability and resiliency for Foundry (Microsoft Learn)
- Agent optimizer overview (Microsoft Learn)
- Publish agents to Microsoft Copilot and Teams (Microsoft Learn)

