AI-103 exam guide: format, skill areas, preparation
Who the AI-103 exam is for
AI-103 is the exam for the Microsoft Certified: Azure AI Apps and Agents Developer Associate certification, an intermediate-level credential. The study guide describes the candidate as an Azure AI engineer who builds, manages and deploys agents and AI solutions that take advantage of Microsoft Foundry. You should have experience developing apps with Python and be familiar with the capabilities of general AI, generative AI and Azure services.
The study guide lists five responsibilities: planning and managing Azure AI solutions, and implementing generative AI and agentic, computer vision, text analysis and information extraction solutions. It adds that in this role you collaborate with business stakeholders, solution architects, data scientists, DevOps engineers and cloud security engineers.
Exam format
| Parameter | Value |
|---|---|
| Exam code | AI-103 |
| Certification | Microsoft Certified: Azure AI Apps and Agents Developer Associate |
| Time to complete | 120 minutes |
| Passing score | 700 or greater |
| Skills measured | Version of April 16, 2026 |
| Languages | English, Chinese (Simplified), Chinese (Traditional), French, German, Japanese, Korean, Italian, Portuguese (Brazil), Spanish |
Conditions worth knowing
- Proctored exam that may include interactive components, according to the exam page.
- Localized versions are updated approximately eight weeks after the English version. If the exam is not available in your preferred language, you can request an additional 30 minutes.
- Preview features: most questions cover generally available (GA) features, but the exam may contain questions on preview features if those features are commonly used.
- Renewal: associate certifications expire annually and can be renewed by passing a free online assessment on Microsoft Learn.
- Retakes: a failed exam can be retaken 24 hours after the first attempt; later retakes follow the retake policy.
Skill areas and weights
The study guide defines five skill areas, each split into groups of objectives.
| Skill area | Weight | Objective groups |
|---|---|---|
| Plan and manage an Azure AI solution | 25–30 % | Choose Foundry services; set up AI solutions in Foundry; manage, monitor and secure AI systems; responsible AI |
| Implement generative AI and agentic solutions | 30–35 % | Build generative applications; build agents; optimize and operationalize generative AI systems |
| Implement computer vision solutions | 10–15 % | Image and video generation; multimodal understanding; responsible AI for multimodal content |
| Implement text analysis solutions | 10–15 % | Language model text analysis; speech solutions |
| Implement information extraction solutions | 10–15 % | Retrieval and grounding pipelines; extracting content from documents |
What each area covers
- Plan and manage: model and service selection, retrieval and indexing choices, deployment options, CI/CD integration, quotas and rate limits, monitoring, managed identity and private networking, safety filters, guardrails and auditing.
- Generative AI and agents: RAG in an application, workflows and multistep pipelines, evaluation, agent roles and tool schemas, function calling and memory, multi-agent orchestration, approval flows, tracing and token analytics.
- Computer vision: generating and editing images and videos, captions and alt text, visual question answering, Content Understanding for visual content, and defences such as detecting indirect prompt injection from embedded text in images.
- Text analysis: extraction of entities and structured JSON, sentiment and sensitive-content detection, translation, and speech to text and text to speech as an agent modality.
- Information extraction: ingestion and indexing, semantic, hybrid and vector search, enrichment skills, OCR and layout analysis, and Content Understanding analyzers.
The two largest areas, planning and managing plus generative AI and agents, together account for the majority of the exam. Time spent in Microsoft Foundry is part of the preparation, not an extra.
How to prepare
- Read the certification page on Microsoft Learn for registration, languages and current conditions.
- Read the AI-103 study guide and copy every objective into a checklist. Check the date of the skills measured at the top of that list.
- Refresh Python and the idea of calling a model, an agent or a tool from a client application by using an SDK.
- In Microsoft Foundry, deploy a model, then build an agent that uses retrieval, a function call and conversation memory. Add tracing and look at what it records.
- Build one retrieval pipeline end to end: ingest documents, index them, compare keyword, vector, hybrid and semantic search on the same questions.
- For each service named in the objectives (Translator, Content Understanding, speech…), note what input it takes and what output it returns.
- Do not skip the operations side: quotas, rate limits, managed identity, private networking, safety filters and safety evaluations make up a whole skill area.
- Try the exam sandbox linked from the certification page to get used to the question formats and the interface.
- Take our AI-103 practice exam under timed conditions, then review each missed question in the free study mode.
- Shortly before the exam, re-check the date of the skills measured on the study guide in case the objectives have changed.
Practice Exam AI-103: Azure AI Apps and Agents Developer Associate
25 free questions with explanations, no account needed.