Training · Azure AI App and Agent Developer Associate

AI-103, building real AI apps, not just describing them.

AI-102 retired June 30, 2026. AI-103 is its actual replacement, an Associate-level, hands-on exam: agents, RAG, computer vision, and information extraction, all built on Microsoft Foundry. Python reading ability is assumed.

What you'll walk away with

Built for developers ready to actually build

If you’ve finished AI-901

Fundamentals gave you the concepts. This is where you actually build and deploy them as working applications.

If you’re exam-bound

Generative AI and agentic solutions alone are 30–35% of this exam, so that’s where most of this course lives.

Curriculum

5 sections, exam-weighted

Module 1 is free to try before you enroll. Everything else unlocks after you enroll.

1. Planning and managing Azure AI solutions

15–20% of exam
1.1 Choosing the right Foundry servicesFree preview
1.2 Planning resource deployment and costLocked
1.3 Managing the AI solution lifecycleLocked

2. Generative AI and agentic solutions

30–35% of exam
2.1 Prompt engineering for real applicationsLocked
2.2 Retrieval-Augmented Generation (RAG)Locked
2.3 Building with the Foundry Agent ServiceLocked
2.4 Multi-agent orchestrationLocked
2.5 Content safety and responsible deploymentLocked

3. Computer vision solutions

10–15% of exam
3.1 Image analysis in a real applicationLocked
3.2 Multimodal models in practiceLocked

4. Text analysis solutions

10–15% of exam
4.1 Language understanding at the API levelLocked
4.2 Summarization and classification pipelinesLocked

5. Information extraction solutions

10–15% of exam
5.1 Document Intelligence in an appLocked
5.2 Grounding a RAG pipeline in real documentsLocked

Capstone

Practical build
Build a working RAG-grounded agent end to end on Microsoft FoundryLocked