Microsoft Certified: Azure AI Apps and Agents Developer Associate

Exam format

Microsoft logo

Microsoft Certified: Azure AI Apps and Agents Developer Associate Exam Format

AI-103

Everything the official exam blueprint specifies: question count, time limit, scoring and the weight each domain carries. This certification validates the ability of a developer to design and implement intelligent applications and AI agents on Microsoft Azure. Candidates are tested across the full AI solution lifecycle, including planning, management, generative AI, agentic workflows, computer vision, text analytics, and information extraction. The exam focuses on applied skills and requires practical experience with Azure AI Foundry, Azure OpenAI, document intelligence, and related AI services. Successful candidates can build production-ready solutions that combine multiple AI capabilities, maintain responsible AI practices, and integrate AI into business processes.

50 questions
120 min
70% to pass
5 domains

Exam facts at a glance

The published specifications for this certification.

SpecificationValue
Questions50
Time limit120 minutes (144s per question)
Passing score70% — 700 or greater (on a scale of 1 to 1,000)
Domains5
LevelAssociate
Question bank here50 questions
Published formatNot published by the provider; exam duration is 120 minutes
SourceOfficial exam specification, published by the vendor

Official domain weights

The percentage of the exam each domain carries. Weight your revision by these numbers, not by what you find hardest.

DomainWeightTopicsPractice Qs
Implement text analysis solutions13%
  • Process text with Azure AI Language
  • Implement language detection and sentiment analysis
  • Extract key phrases and named entities
  • +2 more
6
Implement computer vision solutions13%
  • Analyze images and video with Azure AI Vision
  • Implement object detection, image tagging, and OCR
  • Classify images with custom Azure AI Vision models
  • +2 more
7
Plan and manage an Azure AI solution28%
  • Identify Azure AI services and appropriate use cases
  • Plan security, governance, and responsible AI policies
  • Create and configure Azure AI Foundry projects and hubs
  • +3 more
14
Implement information extraction solutions12%
  • Extract data from forms and documents with Azure AI Document Intelligence
  • Implement prebuilt models for invoices, receipts, and IDs
  • Train custom extraction models for tables and key-value pairs
  • +2 more
6
Implement generative AI and agentic solutions34%
  • Build applications with Azure AI Foundry and Azure OpenAI
  • Integrate Azure AI services using SDKs and APIs
  • Implement retrieval-augmented generation with vector stores
  • +5 more
17

The heaviest domain is Implement text analysis solutions at 13%, and that is where lost points cost you the most. Certilum scores you per domain so you can see the difference between a knowledge gap and a formatting problem.

Note from the vendor record: Successor to AI-102 (Azure AI Engineer Associate), which retired 2026-06-30; AI-103 is in general availability (skills measured as of 2026-04-16). No standalone exam page: /exams/ai-103/ 301-redirects to the certification page. No prerequisites listed; question count not published. Official exam page

More about the Microsoft Certified: Azure AI Apps and Agents Developer Associate exam

Explore the rest of this page series: format, cost, difficulty, requirements and how to prepare.

Exam costPass rateWho can take itStudy plan

Related certifications

Explore other certifications in the same category and see how they stack up.

Microsoft Certified: Intelligent Applications Builder Associate
Microsoft Certified: Intelligent Applications Builder Associate
MicrosoftMicrosoft

Frequently asked questions

Everything you need to know about the Microsoft Certified: Azure AI Apps and Agents Developer Associate exam.