1.1 Current Exam Facts

Why exam facts matter on AI-200

Exam AI-200: Developing AI Cloud Solutions on Azure is not a narrow “prompt engineering” test. It validates that you can ship backend and AI solutions on Azure across the full lifecycle: requirements, design, development, deployment, security, and monitoring. Knowing the official logistics and domain weights early prevents two common failures: studying the wrong Azure AI surface area (for example, treating this like a pure cognitive-services quiz) and under-allocating time to the heaviest domain—Azure data management services for AI.

Treat this section as your source-of-truth checklist. Prefer the official Microsoft Learn study guide and credential pages over third-party summaries whenever numbers conflict. The study guide for AI-200 was updated 2026-05-05; use that date as your freshness signal when you re-check skills measured before exam day.

Credential and exam identity

Official references to bookmark:

• Study guide: https://learn.microsoft.com/en-us/credentials/certifications/resources/study-guides/ai-200
• Credential page: https://learn.microsoft.com/en-us/credentials/certifications/azure-ai-cloud-developer-associate/

Passing at 700/1000 means you need a solid majority across domains—not perfection in one niche. Because the exam can include interactive components, practice reading manifests, portal/CLI-style tasks, and SDK-oriented scenarios, not only flashcard definitions.

The four skills domains (official weights)

Microsoft measures four domains. Memorize both the name and the weight band—exam questions map to these, and your study calendar should too.

Domain 2 sits at 25–30%—the widest band. If your background is “containers only” or “Functions only,” Domain 2 is where you should invest deliberate lab time: embeddings, vector indexes, RU/latency tradeoffs, and retrieval patterns with metadata filters.

How to read weights for study planning

Use midpoints as a rough time budget for a multi-week plan:

1. Domain 1 ≈ 22.5% → containers and orchestration paths
2. Domain 2 ≈ 27.5% → data platforms that back AI features
3. Domain 3 ≈ 22.5% → messaging, events, serverless integration
4. Domain 4 ≈ 22.5% → secrets, config, observability, KQL

That does not mean every session must be proportional day-by-day. It means your weakest heavy domain should never be left for the last weekend. A candidate who can deploy Container Apps but cannot explain Cosmos consistency vs RU cost, or cannot write a basic KQL troubleshooting query, is exposed on two different weight bands.

Proctoring, format expectations, and exam-day behavior

AI-200 is proctored and delivered in English. Plan for a full 120-minute sitting. Interactive items reward calm, systematic reading: identify the resource type, the SDK or control-plane action, and the constraint (security, cost, latency, scale, failure handling).

Practical exam-day habits that map to how Microsoft roles are tested:

• Prefer answers that use the right Azure service for the job (for example, Event Grid for reactive event delivery vs Service Bus for competing consumers and DLQ patterns).
• Prefer answers that respect security and configuration boundaries (Key Vault for secrets; App Configuration for dynamic app settings).
• Prefer answers that close the loop on operations (OpenTelemetry correlation, Application Insights/Azure Monitor, KQL investigation).

Renewal and staying current

The associate credential supports annual free renewal through Microsoft Learn. Passing AI-200 is not a “set and forget” badge—Azure data/AI and container surfaces change. After you pass, schedule the renewal assessment before expiration and skim the latest skills outline whenever Microsoft updates the study guide.

Scenario: choosing what to verify first

Imagine two candidates. Alex memorizes old blog posts that still claim a shorter duration and an outdated domain list. Jordan opens the official study guide (updated 2026-05-05), confirms 120 minutes, 700/1000, and the four domain bands, then builds a study plan with Domain 2 as the largest block. On exam day, Jordan is not surprised by vector/RAG data questions or by Key Vault + KQL items mixed with container deployment scenarios. Alex is. The difference is not intelligence—it is official-fact discipline.

Key definitions to lock before later chapters

• Azure AI Cloud Developer Associate — the credential earned by passing AI-200; signals backend/AI solution skills on Azure, not a generic cloud fundamentals badge.
• Skills measured — the official domain list and bullets; your study checklist should trace back to these, not to random Azure AI marketing pages.
• Interactive components — exam items that may require multi-step configuration-style reasoning beyond a single memorized fact.
• Free annual renewal — Microsoft Learn assessment path that keeps the credential active without retaking the full paid exam every year.

Before you leave this section, you should be able to recite the exam name, credential name, score, duration, language, proctoring note, four domain titles with weight bands, and the two official Learn URLs.