The Microsoft AI-102 exam, titled "Designing and Implementing Azure AI Solutions," targets professionals who build, manage, and deploy AI solutions using Microsoft Azure. This certification replaces the earlier DP-100 certification and reflects Microsoft's consolidated approach to AI solution development on the Azure platform. Candidates who pass this exam earn the Microsoft Certified: Azure AI Solutions Architect designation.
Unlike foundational certifications, AI-102 expects hands-on experience with Azure AI services, programming proficiency in C# or Python, and familiarity with REST APIs, JSON, and containerization concepts. The exam does not test theoretical machine learning knowledge in depth; instead, it focuses on the operational aspects of integrating prebuilt AI capabilities into applications, customizing models with your own data, and managing AI resources at scale.
The certification is ideal for software engineers, data scientists who work with production AI systems, and cloud architects responsible for designing AI-powered solutions. Microsoft recommends at least six to twelve months of hands-on experience with Azure AI services before attempting the exam.
The AI-102 exam follows the standard Microsoft role-based certification format. You can expect between 40 and 60 questions, drawn from a larger pool that Microsoft continuously updates. The total exam duration is 150 minutes, which includes time for case studies that may require you to review a scenario and answer multiple related questions.
| Item | Detail |
|---|---|
| Number of questions | 40-60 |
| Duration | 150 minutes |
| Passing score | 700 out of 1000 |
| Question types | Multiple choice, case studies, drag-and-drop, build list, hot area, review screen |
| Delivery method | Pearson VUE (online proctored or test center) |
| Languages | English, Japanese, Chinese (Simplified), Korean, French, German, Spanish, Portuguese (Brazil) |
Case study questions are particularly important. You will receive a scenario description along with exhibits such as code snippets, architecture diagrams, or Azure portal screenshots. Multiple questions follow the same case study, and you cannot return to a case study once you move past it. Time management matters — allocate roughly two minutes per standard question and slightly longer for case study items.
Flag difficult questions and return to them later. Microsoft exam software allows you to mark questions for review. Answer every question even if you are unsure — there is no penalty for guessing. For case studies, read the scenario thoroughly before looking at the questions. Pay close attention to words like "most," "least," "first," and "best" because they indicate priority or ordering.
Microsoft publishes a detailed document called the "Study Guide" that lists every skill measured on the exam. The percentages below reflect the approximate weight of each domain. Understanding the weighting helps you allocate study time proportionally.
| Domain | Weight |
|---|---|
| Plan and manage Azure AI solutions | 25-30% |
| Implement image and video processing | 15-20% |
| Implement natural language processing | 20-25% |
| Implement knowledge mining | 15-20% |
| Implement conversational AI | 15-20% |
| Implement generative AI | 10-15% |
This domain covers provisioning AI resources in Azure, selecting the appropriate pricing tier, configuring security with virtual networks and private endpoints, managing authentication with keys and managed identities, setting up monitoring and alerts, implementing cost governance with budgets, and ensuring compliance with data residency requirements. You should know how to decide between a single-service and a multi-service Cognitive Services resource and when to deploy containers instead of using the cloud API.
Image and video processing covers Azure AI Vision for image analysis, OCR, and thumbnail generation. It also includes Custom Vision for training classification and object detection models, the Face API for facial detection and recognition, Video Indexer for extracting insights from video content, and Azure AI Document Intelligence for extracting information from forms and documents.
This domain includes Azure AI Language services for text analytics such as sentiment analysis, key phrase extraction, named entity recognition, and language detection. It also covers custom text classification and custom named entity recognition. The Conversational Language Understanding (CLU) service for building intent and entity models is here, along with the Custom Question Answering service for knowledge bases.
Knowledge mining uses Azure AI Search to build search indexes over your data. You must understand index design, data sources, indexers that crawl data automatically, skillsets that add AI enrichment during indexing, and the knowledge store feature that persists enriched content. The domain also includes query syntax, filtering, sorting, scoring profiles, and semantic search capabilities.
Conversational AI focuses on the Bot Framework SDK and Bot Framework Composer. You need to know how to build bots using dialogs, adaptive cards, waterfall dialogs, and language generation. Publishing bots to Azure, connecting them to channels like Microsoft Teams and Slack, and integrating them with Language Understanding and QnA Maker are all exam topics.
This relatively new domain covers Azure OpenAI Service, including model selection, prompt engineering, content filtering, code generation, and responsible AI practices. You should understand how to use GPT models for text generation, the DALL-E model for image generation, and embeddings for semantic search. RAG (Retrieval Augmented Generation) patterns, function calling, and fine-tuning are also part of this domain.
The generative AI domain is the newest addition to the exam. Microsoft has been increasing the weight of this domain as Azure OpenAI adoption grows. Expect scenario-based questions that ask you to choose between a prebuilt AI service and a GPT-based solution for a given requirement.
Before you begin studying for AI-102, ensure you meet these prerequisites:
A structured approach to studying dramatically increases your chance of passing. Here is a recommended study plan organized over a twelve-week period.
| Week | Focus Area | Activities |
|---|---|---|
| 1 | Setup and fundamentals | Create Azure free account, install tools, read exam study guide |
| 2-3 | Plan and manage AI solutions | Provision resources, configure security, set up monitoring |
| 4-5 | Image and video processing | Work with Vision, Custom Vision, Face API, Video Indexer |
| 6-7 | Natural language processing | Use Language, CLU, QnA Maker, custom text services |
| 8 | Knowledge mining | Build search indexes, skillsets, knowledge stores |
| 9 | Conversational AI | Build bots with SDK and Composer |
| 10 | Generative AI | Prompt engineering, Azure OpenAI, RAG patterns |
| 11 | Review and practice tests | Take practice exams, review weak areas |
| 12 | Final preparation | Revisit case studies, schedule exam, rest |
Microsoft provides a variety of free resources for exam preparation. The Microsoft Learn learning path for AI-102 contains multiple modules with interactive exercises that use Azure sandboxes. These modules are the single best starting point because they align directly with the skills measured on the exam.
The official documentation at learn.microsoft.com is comprehensive. Pay special attention to the Azure AI services documentation pages, the REST API reference, and the SDK migration guides. Microsoft also publishes sample code repositories on GitHub under the Azure-Samples organization that demonstrate every major feature.
The Microsoft Tech Community and Microsoft Q&A forums are good places to find answers to specific questions and see what challenges other candidates face.
You need a working environment to follow along with the exercises in this book. The setup involves three steps: an Azure subscription, command-line tools, and an SDK in your programming language of choice.
Go to azure.microsoft.com and click "Start free." You need a Microsoft account (personal or work/school). The free tier includes $200 in credits for the first 30 days, access to more than 55 always-free services, and no obligation to upgrade. Your learning path does not require premium services that exceed the free tier credits.
Azure CLI is the command-line interface for provisioning and managing Azure resources. Download the installer from docs.microsoft.com/cli/azure/install-azure-cli. After installation, open a terminal and run az login to authenticate. The login process opens a browser window where you enter your Azure credentials. Once authenticated, your CLI session can manage resources in your subscription.
Choose C# or Python for the exercises in this book. Both are fully supported by Azure AI services SDKs.
For Python, install the Azure SDK packages using pip:
pip install azure-ai-formrecognizer azure-ai-vision azure-ai-textanalytics azure-cognitiveservices-speech azure-search-documents openaiFor C#, create a new console project and add the NuGet packages through the .NET CLI:
dotnet new console -n Ai102Study
cd Ai102Study
dotnet add package Azure.AI.FormRecognizer
dotnet add package Azure.AI.Vision
dotnet add package Azure.AI.TextAnalytics
dotnet add package Microsoft.CognitiveServices.Speech
dotnet add package Azure.Search.Documents
dotnet add package Azure.AI.OpenAICreate a simple script that lists your Azure subscription. This confirms that your CLI is working and your account has valid credentials.
# Python
from azure.identity import DefaultAzureCredential
from azure.mgmt.resource import SubscriptionClient
credential = DefaultAzureCredential()
client = SubscriptionClient(credential)
for sub in client.subscriptions.list():
print(f"Subscription: {sub.display_name} ({sub.subscription_id})")// C#
using Azure.Identity;
using Azure.ResourceManager;
using Azure.ResourceManager.Resources;
var credential = new DefaultAzureCredential();
var client = new ArmClient(credential);
foreach (var sub in client.GetSubscriptions())
{
Console.WriteLine($"Subscription: {sub.Data.DisplayName} ({sub.Data.SubscriptionId})");
}Your free account credits expire after 30 days. Set up budget alerts in the Azure portal to avoid unexpected charges if you upgrade to a pay-as-you-go subscription. The exercises in this book are designed to stay within the free tier wherever possible.
This book is designed for sequential reading but also serves as a reference for specific topics. Each chapter begins with the learning objectives and ends with a navigation bar to move forward or back. Read the chapters in order during your first pass, then return to individual chapters for targeted review of weak areas.
Every chapter includes at least one hands-on exercise and an exam tip box. Complete the exercises — reading about a service is not enough to build the muscle memory you need for the exam. The code samples are production-quality and you can adapt them for your own projects.
Pay special attention to the tables that compare services, pricing tiers, and feature sets. Exam questions frequently ask you to select the correct service for a given scenario, and knowing the distinguishing features of each service is critical.
Keep the exam study guide PDF open as you read. Cross-reference each topic against the study guide to confirm you are covering everything. Mark off items as you complete them to track your progress.
Register for the exam through the Microsoft Certification dashboard at aka.ms/certification. You can schedule through Pearson VUE for either online proctored or in-person delivery.
Review the study guide and focus on the domains where you scored lowest in the practice test. Do not try to learn new material the night before. Get at least seven hours of sleep. Prepare your ID, ensure your testing area is clean, and install the Pearson VUE secure browser software. Set two alarms — one for waking up and one for the exam start time.
Create a multi-service Cognitive Services resource in the Azure portal. Navigate to "Create a resource," search for "Cognitive Services," and select the multi-service option. Choose your subscription, create a new resource group named rg-ai102-study, select the region closest to you, and choose the Free F0 tier if available or Standard S0. Click "Review + create" and then "Create." After deployment, navigate to the resource and examine the "Keys and Endpoint" section. Copy one of the keys and the endpoint URL — you will use them in the next chapter.
The AI-102 exam tests your ability to design and implement AI solutions using Azure AI services across six domains. Preparation requires a combination of conceptual study and hands-on practice. Set up your Azure environment before diving into the chapters, follow the twelve-week study plan, and complete every exercise in this book. The next chapter begins with a deep dive into Azure AI Services, the foundational building block of every solution you will build on the exam and in real-world projects.