Chapter 4 of 7 · Microsoft AI-901
Generative AI Apps and Agents in Foundry
Writing prompts, deploying a model, a chat client with the Foundry SDK, and creating and calling agents.
- 4topics
- 20 minfast track
- 2diagrams and exhibits
Topics in this chapter
Take them in order, or jump to the one you need.
- 1System and User PromptsAI-901 prompt writing: system, user and assistant messages, clear instructions, context and grounding, few shot examples and…5 min
- 2Deploying a Model in the Foundry PortalAI-901 Foundry portal: resources and projects, deploying a model from the catalog, the chat playground, project endpoint and…Diagram5 min
- 3A Chat Client with the Foundry SDKAI-901 Foundry SDK: azure-ai-projects and azure-identity, AIProjectClient, calling a deployed model, handling secrets, and…5 min
- 4Agents in the Foundry Portal and SDKAI-901 Foundry agents: creating and testing an agent in the portal, instructions and tools, function calling, and a client using…Diagram5 min
About this chapter
What this chapter covers
From here the exam moves from ideas to building. This chapter covers the full path of a generative AI app in Microsoft Foundry: write good prompts, deploy a model and try it in the portal, call it from a small Python client with the Foundry SDK, and build an agent that adds instructions and tools.
You do not need to be an expert programmer, but you must read short Python snippets and know what each part does.
How the topics fit together
The pages follow the order in which you build an app.
- System and user prompts covers how to write instructions that produce reliable answers.
- Deploying a model in the Foundry portal covers creating a project, deploying a model, using the playground and getting the endpoint.
- A chat client with the Foundry SDK covers authenticating, calling a deployed model and keeping secrets safe.
- Agents in the Foundry portal and SDK covers creating an agent, its tools, and a client that talks to it.
Learn the shape of the code, not every line. Create a client from the project endpoint and a credential, then call the model or agent by name, and read the reply from the response.
How to use these pages
Each topic page has the same shape, so you can skim for what you need.
- What it is: the idea in plain words.
- How it works: the mechanism, with examples and tables.
- What to remember for the exam: the facts to memorise and how questions are phrased.
- Common mistakes: the classic traps.
- Check yourself: three questions to test your understanding.
Ready for more practice?
The full Microsoft AI-901 question bank has 511 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by Microsoft. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.