Understand the certification
认证概述
Microsoft Certified: Azure AI Engineer Associate（AI-102）是微软官方颁发的 Azure AI 工程师 Associate 级认证。考试约含 40-60 道题（单选、多选、案例分析、拖放题），考试时长 100 分钟，满分 1000 分，700 分通过，考试费 $165 USD。认证有效期 1 年，需通过年度续期评估保持有效。注意：该认证将于 2026 年 6 月 30 日退役。
考试领域与权重（2025 年 12 月更新版官方技能衡量标准）
- 规划和管理 Azure AI 方案（Plan and manage an Azure AI solution）（20–25%）：选择 Microsoft Foundry Services 服务、部署 AI 模型、监控与安全（密钥管理、身份验证）、Responsible AI 内容安全（Content Safety、Prompt Shields）
- 实现生成式 AI 方案（Implement generative AI solutions）（15–20%）：Microsoft Foundry 平台部署、Azure OpenAI 模型选择与部署、Prompt Flow、RAG 架构（数据接地）、模型微调（Fine-tuning）、DALL-E 图像生成、多模态模型
- 实现 Agentic 方案（Implement an agentic solution）（5–10%）：Microsoft Foundry Agent Service 构建自定义 Agent、Microsoft Agent Framework 复杂工作流、多 Agent 编排与自治能力
- 实现计算机视觉方案（Implement computer vision solutions）（10–15%）：Azure Vision in Foundry Tools（图像分析、OCR、手写识别）、自定义视觉模型（分类与目标检测）、Azure AI Video Indexer 视频洞察提取
- 实现自然语言处理方案（Implement natural language processing solutions）（15–20%）：Azure AI Language（NER、情感分析、PII 检测）、Azure AI Translator、Azure Speech（TTS/STT、SSML）、自定义语言模型与 Question Answering
- 实现知识挖掘和信息提取方案（Implement knowledge mining and information extraction solutions）（15–20%）：Azure AI Search（索引器、Skillset、语义搜索、向量存储）、Azure Document Intelligence（预构建模型 + 自定义模型）、Azure Content Understanding（OCR pipeline、文档分类）
适合人群
- 具备 Python 或 C# 开发经验，希望在 Azure 平台构建端到端 AI 解决方案的工程师
- 正在转型 Azure AI 开发方向，需要 Microsoft 官方认证背书的软件开发者
- 数据科学家和 ML 工程师，希望系统掌握 Azure AI Services 集成与 Responsible AI 治理能力
Exam at a glance
- Exam code
- AI-102
- Provider
- Azure
- Format
- single-choice / multiple-choice
- Language
- English
- Validity
- 3 years
Skills in focus
Learning outcomes
- 通过 AI-102 认证，获得 Microsoft Certified: Azure AI Engineer Associate 徽章
- 掌握在 Microsoft Foundry 平台上端到端规划、部署和管理 AI 解决方案的完整能力
- 能够使用 Azure OpenAI Service 构建生成式 AI 应用、RAG 知识库和 AI Agent
- 熟练集成 Azure AI Language、Vision、Speech 和 Document Intelligence 到实际业务场景
- 掌握 Responsible AI 治理框架，实现 Content Safety 和合规 AI 部署
Who this is for
- Azure 云开发者和 AI 工程师，希望获得微软官方 AI 工程师 Associate 认证
- 有 Python 或 C# 编程经验，正在转向 Azure AI 开发或 AI Agent 构建的软件工程师
- 数据科学家和 ML 工程师，希望系统掌握 Microsoft Foundry 平台和 Azure AI Services 集成能力
Before you start
- 熟悉 Python 或 C# 编程，能够调用 REST API 并处理 JSON 响应
- 了解基本 Azure 服务（建议先取得 AZ-900 Azure 基础认证）
- 具备基础 AI/ML 概念：分类、NLP、计算机视觉原理，以及生成式 AI 基本概念
Your learning path
3 chapters
Plan your preparation
- Learning hours
- 30h
- With relevant experience
- 4-6 weeks
- Starting from scratch
- 8-10 weeks
- Daily study
- 1-2h
Included in your preparation
- 262 道练习题，精准覆盖 AI-102 六大考试域（2025 年 12 月最新版），每题附详细中文解析
- 全新 Agentic AI 专项：Microsoft Foundry Agent Service、多 Agent 编排工作流重点强化
- 生成式 AI 深度覆盖：RAG 架构、Prompt Flow、模型微调、DALL-E 图像生成
- 仿真模拟考试（100 分钟 / 1000 分制），模拟真实 AI-102 考试节奏
- AI 智能辅导，精准识别 Azure AI Services 配置、NLP 场景题和 Responsible AI 治理薄弱点
Preparation tips
- 65 题 90 分钟，平均每题 1 分钟，合理分配时间
- 及格分 70/100，不确定的题先标记跳过，回头再做
- 排除法非常有用 — 先排掉明显错误的选项，剩下的再分析
Common mistakes
- 没有读完所有选项就选答案 — 题目经常有"最佳"答案和"正确但不最佳"的干扰项
- 备考只刷题不理解原理 — 考试场景题需要理解底层概念
- 忽略时间管理 — 在难题上卡太久，导致后面简单题没时间做
Before you register
What is the AI-102 exam format and passing criteria?
The AI-102 exam contains approximately 40-60 questions (single-choice, multiple-choice, case studies, drag-and-drop), with a 100-minute time limit, scored out of 1000 with 700 needed to pass. The exam fee is $165 USD. Certification is valid for 1 year and requires a renewal assessment. The exam is available in English, Simplified Chinese, Japanese, Korean, and other languages.
What core content does the AI-102 exam cover?
The exam covers five domains: Plan and manage Azure AI solutions (25-30%), Implement decision support solutions (10-15%), Implement computer vision solutions (15-20%), Implement natural language processing solutions (25-30%), and Implement generative AI solutions (10-15%). The 2026 update significantly increases Azure OpenAI Service, GPT model integration, and RAG pipeline content.
What prerequisites are needed for AI-102?
Familiarity with Python or C# programming, understanding of REST APIs and JSON, and basic AI/ML concepts are recommended. Having AZ-900 (Azure Fundamentals) first is advisable for Azure service and portal familiarity. Without programming experience, plan an additional 2-4 weeks for Python basics and SDK usage.
What is the difference between AI-102 and AI-100?
AI-102 is the replacement certification for AI-100. AI-102 adds Azure OpenAI Service, Document Intelligence (formerly Form Recognizer), Cognitive Search with RAG integration, and other modern AI technologies. The exam format has also been upgraded to 100 minutes with a 1000-point scale. AI-100 was retired in 2021 and is no longer available.
Can these practice questions help me pass the AI-102 exam?
This set of 262 practice questions covers core AI-102 exam domains with detailed explanations and mock exams. Combined with the AI tutoring system, it precisely identifies weak areas for targeted improvement. We recommend using it alongside Microsoft Learn official learning paths and hands-on labs, and achieving a consistent 85%+ accuracy rate before taking the actual exam.
What is the career value of AI-102 certification?
AI-102 is the most important Associate-level certification in Azure AI development, suitable for AI engineers, data scientists, and cloud developers. Certified professionals typically earn in the $120K-$180K annual salary range. This certification is also a recommended prerequisite for advanced certifications like AB-100 (Agentic AI Architect), helping advance your career path.
Learner experiences
Loading reviews…
Done reading? Next steps
Start with the first chapter.
2 free chapters · Full preparation from $29