Cert-prep courses with hands-on practice for AWS, Azure, Kubernetes, security and AIModule 1 of every course free, no card →
CloudaQube

AI-103exam prep · Microsoft · intermediate

Microsoft AI-103: Developing AI Apps and Agents on Azure

Prepare for the Microsoft AI-103 exam (Azure AI Apps and Agents Developer Associate), the successor to AI-102. Learn to plan, secure and manage Microsoft Foundry solutions, build generative apps and agents with the Foundry SDK, and implement vision, text, speech, retrieval and document extraction solutions with Azure AI Search and Content Understanding. Text-first lessons with runnable Python and REST examples, two hands-on labs, module quizzes, a capstone project and a long-form practice exam, aligned to the skills measured updated April 16, 2026.

  • 25 hours
  • 47 lessons
  • 6 modules
  • 50-question practice exam
  • updated October 2026
Module 1 freeNo credit cardFull course from $8.99/mo
exam coverage

Mapped to the AI-103 exam guide

Every module is tagged with the exam domain it covers.

—Plan and manage an Azure AI solutionModule 1
—Implement generative AI and agentic solutionsModule 2
—Implement computer vision solutionsModule 3
—Implement text analysis solutionsModule 4
—Implement information extraction solutionsModule 5
—Exam readiness and reviewModule 6
try it

Try it before you sign up

Read a full lesson and answer 3 questions from the Module 1 quiz. No account needed.

free lesson · read · 30 min

Choosing Models and Foundry Services for Generative AI and Agents

AI-103 is Developing AI Apps and Agents on Azure. Passing it earns the Microsoft Certified: Azure AI Apps and Agents Developer Associate certification. The exam replaced AI-102, which Microsoft retired on June 30, 2026,…
question 1 of 3 · from the Module 1 quiz
A team must build a vector index over its company documents. Which kind of model creates the vectors?
Pick an answer.
curriculum

6 modules, 47 lessons

Module 1 is free — lessons, hands-on walkthrough and quiz. The course ends with a 50-question practice exam.

  • readChoosing Models and Foundry Services for Generative AI and Agents30 min
  • videoVideo: Plan and prepare to develop AI solutions on Azure30 min
  • readSetting Up AI Solutions in Foundry: Infrastructure, Deployments and CI/CD30 min
  • videoVideo: Select, deploy, and evaluate Microsoft Foundry models29 min
  • readManaging, Monitoring and Securing AI Systems30 min
  • videoVideo: Securing AI with Azure AI Foundry11 min
  • readResponsible AI Across Generative and Agentic Systems30 min
  • videoVideo: Implement a responsible generative AI solution in Microsoft Foundry21 min
  • quizQuiz: Plan and Manage an Azure AI Solution18 min
how it's built

Built by the CloudaQube team

No celebrity instructor — here is exactly what goes into the course instead.

  • 01 Follows the official AI-103 exam guide — every module is tagged with the exam domain it covers.
  • 02 3 guided hands-on walkthroughs you run in your own Azure subscription.
  • 03 5 module quizzes and a 50-question practice exam — every answer comes with an explanation.
  • 04 Last updated October 2026.
  • 05 Any module can become a hands-on lab in Labs Studio.
outcomes

What you'll be able to do

  • 01 Explain how Microsoft Foundry resources, projects, models, Foundry Tools and agents fit together and choose models, deployment types and services for a requirement
  • 02 Identify the quota, monitoring, identity, networking and guardrail settings that make a Foundry solution secure, observable and affordable
  • 03 Build generative applications with the Foundry SDK, including retrieval-augmented generation, structured output and multistep pipelines
  • 04 Build agents with function calling, knowledge tools, memory and multi-agent orchestration, protected by approvals and least-privilege tool access
  • 05 Apply evaluators, tracing and error analysis to detect fabrications, measure quality and safety, and gate releases in a CI/CD pipeline
  • 06 Implement image and video generation, multimodal understanding, text analysis, translation and speech solutions with responsible AI controls
  • 07 Evaluate retrieval and extraction designs using Azure AI Search hybrid and agentic retrieval and Azure Content Understanding analyzers
before you start
  • — Python programming experience (the exam assumes you build apps in Python)
  • — Familiarity with Microsoft Foundry basics, for example from the CloudaQube AI-901 course (microsoft-ai-901-azure-ai-fundamentals)
  • — An Azure subscription where you can create a Foundry resource and deploy models (needed for the hands-on labs)
  • — Comfort with REST APIs, JSON and the Azure CLI
questions

Before you start AI-103

Is this course current for the AI-103 exam?

It follows the current AI-103 exam guide — all 6 exam domains, with the weights shown on this page, and was last updated October 2026.

How do I book the AI-103 exam, and what does it cost?

You book it directly with Microsoft. The official exam page lists the current fee, format, length and passing score, which can vary by country.

Is this an exam dump?

No. The lessons and quiz questions are written to teach the exam objectives, and every answer comes with an explanation. Real exam questions (“dumps”) break the exam provider’s candidate agreement and can cost you the certification.

What do I need for the hands-on walkthroughs?

An Azure subscription — a free account works for most walkthroughs. Delete what you create when you finish so you don’t keep paying.

How long does it take?

About 25 hours of lessons, videos and hands-on walkthroughs across 6 modules, plus a 50-question practice exam.

What’s free?

Module 1 in full — lessons, hands-on walkthrough and quiz — with no card required. Any paid plan (from $8.99/mo) unlocks all 6 modules and the practice exam, and starts with a 14-day free trial.

get started

Start AI-103 with Module 1 free.

No card required. Unlock every module and the practice exam from $8.99/mo when you’re ready.

See pricing