Skip to content
XUtilHub
← All courses

Microsoft · AI-103 · Intermediate

AI-103: Developing AI Apps and Agents on Azure — free course

A free, complete course for the AI-103 exam (Azure AI Apps and Agents Developer Associate). Plan, secure and monitor Microsoft Foundry solutions, build RAG apps and agents, multi-agent workflows, evaluation and observability, plus vision, text, speech and information extraction, with Python examples.

Exam
AI-103
Certification
Microsoft Certified: Azure AI Apps and Agents Developer Associate
Level
Associate (intermediate)
Passing score
700 out of 1000
Skills outline
As of April 16, 2026
Course length
18 lessons · about 4 hours

What the exam measures

25–30%

Plan and manage an Azure AI solution

  • Choose models and Foundry services
  • Set up Foundry, deployments and CI/CD
  • Manage quotas, monitoring and security
  • Guardrails, safety evals and governance

30–35%

Implement generative AI and agentic solutions

  • Generative apps with the Foundry SDK
  • Retrieval-augmented generation (RAG)
  • Evaluate models and apps
  • Build agents with tools and memory
  • Multi-agent workflows and approvals
  • Tuning, reflection and observability

10–15%

Implement computer vision solutions

  • Image and video generation
  • Multimodal understanding and visual safety

10–15%

Implement text analysis solutions

  • Text analysis and translation
  • Speech and voice agents

10–15%

Implement information extraction solutions

  • Retrieval and grounding pipelines
  • Document extraction with Content Understanding

Course syllabus · 18 lessons · about 4 hours

Your progress

0 of 18 lessons complete · saved in this browser

Start the course →

  1. Start here

    What the exam covers, how it differs from AI-102, and a study plan.

    1. 1AI-103 exam guide and study plan▶ Video9 min

  2. Plan and manage an Azure AI solution

    25–30% of the exam

    Choose models and Foundry services, set up and deploy, then manage cost, monitoring, security and responsible AI.

    1. 2Choose models and Foundry services▶ Video14 min
    2. 3Set up Foundry, deployments and CI/CD▶ Video14 min
    3. 4Manage quotas, monitoring and security14 min
    4. 5Guardrails, safety evals and governance13 min

  3. Implement generative AI and agentic solutions

    30–35% of the exam

    Generative apps with the Foundry SDK, RAG, evaluation, agents, multi-agent workflows, and tuning and observability.

    1. 6Generative apps with the Foundry SDK14 min
    2. 7Retrieval-augmented generation (RAG)15 min
    3. 8Evaluate models and apps13 min
    4. 9Build agents with tools and memory16 min
    5. 10Multi-agent workflows and approvals15 min
    6. 11Tuning, reflection and observability14 min

  4. Implement computer vision solutions

    10–15% of the exam

    Generate and edit images and video, multimodal understanding, and responsible AI for visual content.

    1. 12Image and video generation12 min
    2. 13Multimodal understanding and visual safety15 min

  5. Implement text analysis solutions

    10–15% of the exam

    Language model text analysis, translation, and speech for agents.

    1. 14Text analysis and translation14 min
    2. 15Speech and voice agents13 min

  6. Implement information extraction solutions

    10–15% of the exam

    Retrieval and grounding pipelines with Azure AI Search, and document extraction with Content Understanding.

    1. 16Retrieval and grounding pipelines15 min
    2. 17Document extraction with Content Understanding13 min

  7. Final review

    Everything on one page before the exam.

    1. 18Cheat sheet and exam-day tips10 min

AI-103 practice test

85 exam-style questions with explanations and a score by skill area.

Start practice test →

Frequently asked questions

Is this AI-103 course free?

Yes. Every lesson, quiz and the practice test are free, with no sign-up. Your progress is saved in your browser.

Does AI-103 replace AI-102?

Yes. AI-103 is the associate exam for Azure AI engineers who build apps and agents on Microsoft Foundry. Compared with AI-102 it puts far more weight on generative AI, agents, multi-agent workflows, evaluation and observability, and less on single-purpose services.

Should I pass AI-901 before AI-103?

It isn't required, but it helps. AI-901 teaches the concepts (responsible AI, models, deployments, Foundry basics); AI-103 expects you to build, secure and operate real solutions in Python. Our free AI-901 course is a good warm-up.

How much coding is on AI-103?

A lot of reading code. The audience profile asks for experience developing apps with Python. Expect questions that show SDK snippets, REST calls or configuration and ask what is missing or what happens.

How long does it take to prepare for AI-103?

With some Azure and Python experience, plan for four to eight weeks: one pass through the lessons, hands-on labs in a Foundry project, then practice tests until you are comfortably above the pass mark.

Is this an official Microsoft course?

No. XUtilHub is independent and not affiliated with Microsoft. The course follows the official study guide, which you should check for the latest changes before booking your exam.

XUtilHub is independent and not affiliated with or endorsed by Microsoft. This course follows the public study guide; check the official exam page for the latest changes before you book.