Microsoft Certified: Azure AI Cloud Developer Associate

 

---

Before You Start Studying

Before diving into the study sections, review the prerequisites and recommended background knowledge below to understand what will help you succeed.

 

Recommended Exam Path

AI-200 is a new 2026 exam and the successor to AZ-204: its official course (AI-200T00) replaces the retired AZ-204T00. It tests back-end developers who build AI-powered cloud solutions on Azure.

If you are new to Azure, start with AZ-900: Azure Fundamentals or AI-901: Azure AI Fundamentals first. If you hold (or studied for) AZ-204, you already know a large part of this exam: containers, Functions, messaging, Key Vault, and App Configuration carry over. What is genuinely new is the vector database content (Cosmos DB, PostgreSQL, Managed Redis).

Python matters here: every official learning path for AI-200 lists Python as the primary language, and code questions use the Python Azure SDKs.

 

Prerequisites

There are no strict prerequisites, but the exam assumes you can:

• Read and write Python comfortably
• Work with Azure SDKs and REST-based services
• Understand containers (images, registries, deployment)
• Reason about messaging and event-driven patterns

 

Recommended Background Knowledge

Containerized Applications

• Building and tagging container images
• What a container registry does
• The difference between App Service, Container Apps, and AKS

Vector Databases and AI Patterns

• What embeddings are and why dimensions matter
• Similarity search (exact vs approximate nearest neighbor)
• The retrieval-augmented generation (RAG) pattern

Messaging and Eventing

• Queues vs topics vs events
• Why dead-letter queues exist

Monitoring

• What distributed tracing is
• Reading basic log queries

 

---

Step-by-Step Study Guide

 

Step 1: Review the Official Study Guide

What to do:

• Open the official AI-200 study guide
• Read the skills measured sections (updated 2026-05-05)
• Note any topics that are new to you
• Use it as your checklist throughout your prep

Link to The Official Study Guide

 

Step 2: Schedule Your Exam

What to do:

• Choose a date that gives you enough time for study and practice
• Schedule through the official Microsoft certification page
• Put the date on your calendar and plan backwards

Recommended timing:

• Coming from AZ-204 or daily Azure development: 3 to 5 weeks (focus on the data services area)
• New to Azure development: 6 to 10 weeks

Certification and Exam Details Page

 

Step 3: Go Through the Official Learning Paths

What to do:

• Complete the official AI-200T00 course content (9 learning paths)
• Work in course order: containers, Cosmos DB, PostgreSQL, Redis, integration, secrets and configuration, observability
• Take notes on concepts you cannot explain in simple terms

Official Course Page (AI-200T00)

Spend disproportionate time on the data services area (25-30% of the exam). It has the least overlap with AZ-204 and leans on documented hard boundaries: vector index dimension limits, immutable vector policies, Redis modules that must be enabled at creation, and scale-to-zero rules. New exams test exactly these documented constraints.

 

Step 4: Get Hands-On Practice

AI-200 rewards real experience building AI back ends.

You should aim to get experience with:

• Building and deploying a container to Container Apps or App Service
• Storing embeddings and running a vector similarity query in at least one store
• Processing a queue with Service Bus and handling dead-lettered messages
• Retrieving a secret from Key Vault with a managed identity
• Following one request through a distributed trace

A great way to build this experience is the official end-to-end RAG sample, which combines PostgreSQL with pgvector, Azure OpenAI, and Container Apps in one deployable Python project:

RAG on PostgreSQL sample (Azure-Samples)

 

Step 5: Benchmark Your Knowledge

What to do:

• Use CertiAce to benchmark your readiness across all four skill areas
• Practice case studies and time management
• Return to the learning paths for any unstable topic

Note on the official practice assessment: Microsoft has not published one for AI-200 yet. They usually appear within 8 weeks of an exam leaving beta, so check the certification page periodically.

CertiAce AI-200 Exam Practice

 

Step 6: Take the Exam

The day before:

• Review your weak topics only
• Revisit the hard boundaries (index limits, tier constraints, scaling rules)
• Avoid learning brand new topics

On exam day:

• Read questions carefully and identify what they are truly asking
• Eliminate wrong options first
• Watch for wording that implies constraints such as least privilege, cost, scale to zero, or "must not store credentials"

 

---

Additional Learning Resources

 

Vector Databases and AI Data Services

• Vector search in Azure Cosmos DB for NoSQL: the current reference for the three vector index types and their limits
• Cosmos DB vector search performance tips: choosing between flat, quantizedFlat, and diskANN
• DiskANN on Azure Database for PostgreSQL now generally available
• Redis vector search concepts: FLAT vs HNSW with query syntax
• RedisVL documentation: the Python client used in the Azure Managed Redis tutorials
• Azure Cosmos DB video library

 

Containers and Serverless

• KEDA documentation: Scaling Deployments: Container Apps exposes KEDA scaler config almost verbatim
• AKS Engineering Blog: current posts on running AI workloads on AKS
• Azure Functions University: free lesson-based Functions curriculum with Python coverage

 

Messaging

• Choose between Event Grid, Event Hubs, and Service Bus: the single page that settles service-selection questions

 

KQL and Monitoring

• Must Learn KQL: the best-known free KQL curriculum
• Kusto Detective Agency: gamified KQL challenges that build real query fluency

 

Other Resources

• Exam sandbox: demo the exam UI and question types
• AI-200 launch announcement (Microsoft Skills Hub)