Exam Strategy

AI-200 vs AI-102: Which Azure AI Exam Should You Take First?

Both exams carry the Azure AI Associate badge — but they test completely different skills. Here is how to decide which to take first and whether you should pursue both.

The one-line summary

AI-102 asks: "Can you configure and use Azure's AI services?" AI-200 asks: "Can you build and run the cloud infrastructure that AI applications live on?" If you write application code and deploy to Azure, AI-200 is your exam. If you configure Cognitive Services and AI Search for non-developer roles, AI-102 is yours.

What AI-102 covers

The Azure AI Engineer Associate exam (AI-102) tests your ability to:

  • Create and configure Azure AI Services (Vision, Language, Speech, Document Intelligence)
  • Build solutions with Azure OpenAI Service — prompt engineering, completions, embeddings
  • Set up Azure AI Search including indexers, skillsets, and semantic ranking
  • Implement responsible AI principles
  • Secure and monitor AI services (mostly configuration, not infrastructure)

The audience is AI engineers, data scientists, and solution architects who consume Azure AI services rather than build the platform underneath them.

What AI-200 covers

The Azure AI Cloud Developer Associate exam (AI-200) tests your ability to:

  • Containerise AI workloads — ACR, AKS, Container Apps, Dapr
  • Store AI data — Cosmos DB with DiskANN vector search, PostgreSQL + pgvector, Azure Cache for Redis
  • Build event-driven pipelines — Service Bus, Event Grid, Event Hubs, Azure Functions
  • Secure and operate — Key Vault, managed identity, RBAC, Azure Monitor, KQL

The audience is developers and platform engineers who build and deploy AI-powered applications on Azure infrastructure.

Difficulty comparison

Both are associate-level, meaning roughly equivalent difficulty. From my experience preparing for AI-200:

  • AI-102 has more breadth (many Azure AI services to learn) but less depth per service
  • AI-200 has less breadth but goes deeper — you need to know the internals of AKS networking, Cosmos DB partition strategy, pgvector index types, and Service Bus session semantics
  • AI-200 is harder if you lack a containers or databases background; easier if you already hold AZ-204

Which to take first

My recommendation depends on your background:

  • Backend / cloud developer: AI-200 first. It extends your existing knowledge of containers and APIs into AI-native patterns. AI-102 can follow as a complement.
  • Data scientist / ML engineer: AI-102 first. It directly applies to your daily work with models and AI services. AI-200 teaches you how to operationalise those models at scale.
  • Solutions architect: Take both in parallel (6–8 weeks total) — they complement each other well and give you the full Azure AI stack.
  • Career switcher with no Azure background: AZ-900 → AI-102 → AI-200. Each builds on the last without prerequisite gaps.

Earning both certifications

There is no formal prerequisite — you can take them in any order. Many candidates who hold both report that the combination opens doors to senior roles that require both AI engineering skills and cloud platform depth. The certifications do not stack into a higher tier, but on a CV the combination is visible and distinctive.

Time estimate to earn both from scratch (with no prior Azure experience): 3–4 months studying 1–2 hours per day.

Frequently asked questions

What is the difference between AI-200 and AI-102?

AI-102 (Azure AI Engineer Associate) is about consuming and configuring pre-built Azure AI Services (Azure OpenAI, Cognitive Services, AI Search). AI-200 (Azure AI Cloud Developer Associate) is about building and deploying the infrastructure that hosts those AI applications — containers, databases, messaging, security, and observability.

Can I take AI-200 without any Azure experience?

Technically yes, but it is harder. AI-200 assumes comfort with cloud-native concepts: containers, AKS, managed databases, event-driven architecture. If you are new to Azure, completing AZ-900 (Fundamentals) first is worth the two weeks of study.

Do AI-200 and AI-102 overlap?

Yes, but only in Azure Identity and security fundamentals. AI-102 does not cover AKS, Cosmos DB vector search, Service Bus, or Event Hubs. AI-200 does not cover Cognitive Services configuration, Azure AI Search indexers, or model fine-tuning.

Which exam pays better?

Both are associate-level certifications. Roles that benefit from AI-200 (cloud-native developer, platform engineer) tend to align with higher base salaries than AI-102-focused roles (AI solution architect, AI engineer) — but both are strong complements.