Certification guide
Microsoft’s AI-200 exam, Azure AI Cloud Developer Associate, is the direct successor to the retiring AZ-204 Azure Developer track. It shifts the associate-tier Azure developer credential from generic compute-and-storage patterns toward the concrete work of shipping AI-integrated applications: containerized microservices, vector-backed data stores, event-driven service integration, and end-to-end observability. If you build applications on Azure that call foundation models, retrieve context from vectors, or orchestrate AI workflows across services, this is the developer credential aligned to that work.
Because AI-200 is still in Microsoft’s beta window through mid-2026, the exam blueprint reflects the current state of Azure AI: containers-first deployment, vectors and semantic retrieval as the default data pattern, managed identity as the security answer, and OpenTelemetry-based observability. The practice bank on this page mirrors that reality across every question rather than pattern-matching to older Azure-developer material.
Containerized AI solutions on Azure
Domain 1 · 23%Deploy and operate AI workloads on Azure Container Registry, Azure Container Apps, AKS, and App Service. Covers ACR Tasks with base-image triggers and geo-replication, Container Apps multi-revision canaries and KEDA scale-to-zero rules, Azure Workload Identity federation to Kubernetes service accounts, tag immutability, and managed-identity image pulls.
AI data management services
Domain 2 · 28%Design AI data layers on Cosmos DB for NoSQL (session consistency, tenant partitioning, diskANN vector indexes, change feed), Azure Database for PostgreSQL Flexible Server with pgvector (HNSW indexes, cosine ops, Memory-Optimized sizing, PgBouncer connection pooling), and Azure Managed Redis (cache-aside, FT.CREATE VECTOR indexes, volatile-lru eviction). The heaviest domain by weighting.
Azure services integration
Domain 3 · 23%Wire AI workloads into Azure Service Bus (PeekLock, sessions, correlation filters, DLQ), Event Grid (at-least-once delivery, subjectBeginsWith filters, dead-lettering), and Azure Functions (Cosmos DB triggers on the change feed, Premium plan for VNet + no cold starts, Durable Functions fan-out/fan-in, output bindings). Emphasises native bindings over hand-written client code.
Security, monitoring, and troubleshooting
Domain 4 · 25%Secure and observe AI applications: Key Vault references from App Configuration, SecretNearExpiry rotation via Event Grid, DefaultAzureCredential chain, Azure RBAC vs vault access policies, soft delete recovery, user-assigned managed identity sharing, and Application Insights via the Azure Monitor OpenTelemetry Distro, including adaptive sampling, Live Metrics streams, KQL for error triage, and Standard availability tests.
The Certifym practice bank for AI-200 is 1,020 original questions across 17 industry-vertical sets, generic enterprise, healthcare, financial services, manufacturing/OT, federal (FedRAMP-High-authorized services only), retail, SaaS, higher education, energy, telecom, media, insurance, legal, non-profit, gaming, automotive, and real estate, with a 70% pass mark on the master simulator. That is deliberately calibrated as an honest raw-score equivalent of Microsoft’s 700/1000 scale: passing here means you have earned it on the heavy Domain 2 vector-data questions and the situational security-and-observability judgement calls, not luck in the lighter container-lifecycle items.
Microsoft Azure AI Cloud Developer - Practice Exam
Certifym AI-200 master simulator drawing from a large question bank. Each attempt serves 60 randomly selected questions across all four domains and every industry vertical. Version…
Subscribe to startFrequently asked questions about AI-200
What is the Microsoft AI-200 exam?
AI-200 is Microsoft’s Azure AI Cloud Developer Associate exam, the direct successor to the retiring AZ-204 Azure Developer track. It shifts the associate-tier Azure developer credential away from generic compute-and-storage patterns toward the work of shipping AI-integrated applications: containerized microservices, vector-backed data stores, event-driven service integration, and end-to-end observability.
How many questions are on the AI-200 exam?
Roughly 40 to 60 questions. The format is multiple choice plus case study, so some items are standalone and others are grouped around a scenario you read once and then answer several questions about.
How long is the AI-200 exam?
About 100 minutes. With roughly 40 to 60 items in that window, the case-study sections are where time pressure actually shows up, budget reading time for the scenario before you start answering.
What is the passing score for AI-200?
700 out of 1000 on Microsoft’s scaled scoring system. That is a scaled score rather than a raw percentage, so it does not map cleanly to “70% of questions correct”: Microsoft adjusts for item difficulty across exam forms.
Is AI-200 still in beta?
Yes. This page tracks version v1-2026-05, and AI-200 remains in Microsoft’s beta window through mid-2026. Because of that, the blueprint reflects the current state of Azure AI rather than older Azure-developer material: containers-first deployment, vectors and semantic retrieval as the default data pattern, managed identity as the security answer, and OpenTelemetry-based observability.
What domains does AI-200 cover, and how are they weighted?
Four domains: Containerized AI solutions on Azure (23%), AI data management services (28%), Azure services integration (23%), and Security, monitoring, and troubleshooting (25%). AI data management services is the heaviest domain by weighting, so vector-backed data stores deserve the most study time.
Who is AI-200 for?
Developers who build applications on Azure that call foundation models, retrieve context from vectors, or orchestrate AI workflows across services. It is the associate-tier developer credential aligned to that work, and the blueprint assumes you are already deploying containers, wiring up messaging and serverless services, and instrumenting applications rather than learning those things for the first time.
How hard is the AI-200 exam?
The difficulty concentrates in two places: the heavy Domain 2 vector-data questions: Cosmos DB diskANN indexes, pgvector HNSW and cosine ops, Redis FT.CREATE VECTOR indexes, and the situational security-and-observability judgement calls in Domain 4, where several answers look defensible. The container-lifecycle items in Domain 1 are comparatively lighter.
How do I prepare for AI-200?
Work the blueprint in weight order and drill scenarios rather than memorising service names. Certifym publishes 1,020 original AI-200 questions across 17 industry-vertical sets, generic enterprise, healthcare, financial services, manufacturing/OT, federal, retail, SaaS, higher education, energy, telecom, media, insurance, legal, non-profit, gaming, automotive, and real estate, with a 70% pass mark on the master simulator, deliberately calibrated as an honest raw-score equivalent of Microsoft’s 700/1000 scale.
What is the difference between AI-200 and AZ-204?
AI-200 is the direct successor to the retiring AZ-204 Azure Developer track. AZ-204 tested generic compute-and-storage development patterns; AI-200 rebuilds the same associate developer tier around AI-integrated applications, containers, vector data stores, event-driven integration, and observability. If you are choosing between them now, AI-200 is the forward-looking credential.
Trademark notice & independence. Certifym.net is operated by Certifym Exam Services, LLC and is not affiliated with, endorsed by, or sponsored by Microsoft Corporation. Microsoft, Azure, and all related marks are trademarks of Microsoft Corporation, used here only for factual reference to the exam.
All practice questions, explanations, and study materials on this page are original content produced by Certifym Exam Services, LLC. They are not sourced from actual Microsoft exam questions and are not endorsed by Microsoft.