Study Strategies and Exam Tips for the AI-102

Building Your Study Plan

The most effective approach to the AI-102 is a structured study plan that allocates time proportionally to domain weights and includes substantial hands-on practice:

Week-by-Week Study Schedule

Week 1: Azure AI Fundamentals and Solution Planning

• Study Domain 1: Azure AI services overview, resource provisioning, security
• Learn Azure AI Foundry portal navigation and resource management
• Understand RBAC, managed identities, and Key Vault integration
• Set up an Azure free account and create your first AI service resources
• Take Domain 1 practice questions

Week 2: Content Moderation and Computer Vision Foundations

• Study Domain 2: Azure AI Content Safety — text and image moderation
• Begin Domain 3: Azure AI Vision — Image Analysis 4.0 API
• Practice with Vision Studio for image analysis, OCR, and spatial analysis
• Build a simple image analysis application using the Python SDK
• Take Domain 2 and early Domain 3 practice questions

Week 3: Advanced Computer Vision and Face API

• Continue Domain 3: Custom Vision, Face API, Video Indexer
• Train a custom image classification model in Custom Vision portal
• Implement face detection and verification using the Face SDK
• Practice OCR with Document Intelligence for structured documents
• Take Domain 3 practice questions

Week 4: Natural Language Processing Deep Dive

• Study Domain 4: Azure AI Language — sentiment, NER, key phrases, PII
• Build CLU (Conversational Language Understanding) models
• Implement custom text classification and custom NER
• Practice with Language Studio for interactive testing
• Take early Domain 4 practice questions

Week 5: NLP Completion, Knowledge Mining, and Document Intelligence

• Continue Domain 4: Azure AI Speech, translation, question answering
• Study Domain 5: Azure AI Search — indexing, skillsets, knowledge stores
• Study Domain 5: Azure AI Document Intelligence — prebuilt and custom models
• Build an end-to-end knowledge mining pipeline with AI enrichment
• Take Domain 4 and 5 practice questions

Week 6: Generative AI, Review, and Practice Exams

• Study Domain 6: Azure OpenAI Service — models, deployments, prompt engineering
• Implement RAG pattern with Azure AI Search and Azure OpenAI
• Review all six domains, focusing on weak areas identified in practice tests
• Take 2-3 full-length practice exams under timed conditions
• Review every incorrect answer and the associated Microsoft Learn documentation

Free Study Resources

Hands-On Practice Is Essential

The AI-102 is fundamentally different from fundamentals exams like the AZ-900 or AI-900. You must have hands-on experience with:

Must-Practice Skills

1. Creating and configuring Azure AI service resources using the Azure portal and Azure CLI
2. Calling Azure AI REST APIs — constructing HTTP requests with correct headers, endpoints, and payloads
3. Using Azure AI SDKs — writing Python or C# code to interact with Azure AI services
4. Training custom models — Custom Vision, CLU, custom NER, custom text classification
5. Building AI Search indexes — creating data sources, indexers, skillsets, and indexes
6. Deploying Azure OpenAI models — provisioning deployments and calling the completions/chat API
7. Implementing content filtering — configuring Azure AI Content Safety and content filters

Recommended Lab Exercises

• Deploy an Azure AI Vision resource and analyze images using the SDK
• Build a Custom Vision image classification project and publish the model
• Create a CLU application with intents, entities, and utterances
• Set up an Azure AI Search index with cognitive skills enrichment
• Deploy GPT-4o in Azure OpenAI Service and implement a chat completion endpoint
• Configure Azure AI Content Safety for text and image moderation

Exam-Day Strategies

Time Management

With 40-60 questions in 120 minutes, you have approximately 2-3 minutes per question. Case study sections require more time.

Strategy:

1. Case studies first (if presented): Spend 15-20 minutes per case study section — read the scenario carefully before answering questions
2. First pass (60-70 minutes): Answer all non-case-study questions you are confident about. Flag anything taking more than 2 minutes.
3. Second pass (30-40 minutes): Return to flagged questions with fresh perspective
4. Final review (10 minutes): Quick scan of all flagged answers

Key Patterns in AI-102 Questions

"Which service should you use?" questions: Match the Azure AI service to the specific capability described. Know the boundaries between services (e.g., Azure AI Language for NER vs. Azure AI Document Intelligence for form field extraction).

"What code should you write?" questions: You may see code snippets with a blank line or dropdown. Know the SDK class names, method signatures, and common parameters for each service.

"How do you configure?" questions: Understand Azure portal configuration, ARM templates, and Azure CLI commands for provisioning AI resources.

"What is the correct order?" questions: Sequence the steps for building a solution (e.g., create resource → create project → add training data → train model → publish model → call endpoint).

Common Traps to Avoid