Exam overview — AI-200
Exam AI-200: Developing AI Cloud Solutions on Azure leads to the Microsoft Certified: Azure AI Cloud Developer Associate credential. It targets developers who build, integrate, secure, and monitor AI solutions on Azure, with an emphasis on back-end services and components.
At a glance
| Credential | Azure AI Cloud Developer Associate |
| Length | ~120 minutes |
| Pass score | 700 / 1000 (scaled) |
| Question styles | Multiple choice, multiple response, case studies, drag-and-drop, possibly labs |
| Renewal | Annually, via a free Microsoft Learn assessment |
Freshly out of beta
Microsoft hasn't yet published a dated "Skills measured as of…" heading or a change log for AI-200, and the official practice assessment isn't available yet either — both usually appear within about eight weeks of an exam going generally available. Re-check the study guide before you sit it.
Skills measured — the four domains
What you should already be proficient in
- Azure SDKs and third-party SDKs used on Azure
- Azure data management services (Cosmos DB, PostgreSQL, Redis)
- Azure monitoring and troubleshooting (Azure Monitor, Log Analytics, KQL)
- Azure messaging and eventing (Service Bus, Event Grid)
- Vector databases and embedding-based retrieval
- Python programming — this is what the official study guide lists. The exam itself is language-agnostic in how it scores you; only code snippets differ. The questions and snippets on this site are written in C# / .NET, so read the Python prerequisite as "you can read code in the SDK of your choice".
- Implementing containerized applications on Azure (ACR, Container Apps, AKS)
Key cross-cutting themes
These ideas show up across multiple domains — make sure they are second nature:
- Managed identity + RBAC over stored secrets — for ACR pulls, Key Vault, Cosmos, etc.
- Vector search / RAG — the same embedding-and-retrieve pattern recurs in Cosmos DB, PostgreSQL/pgvector, and Redis.
- Event-driven scaling — KEDA in Container Apps, queue/topic triggers in Functions.
- Observability — OpenTelemetry tracing into Application Insights, analyzed with KQL.
Ready? Start with Domain 1 or jump to the study plan.