- What "AI-102" Actually Refers To Today
- The Retirement That Changed Everything
- What the Current Exam Actually Tests
- Domain-by-Domain Breakdown
- Exam Mechanics: Fee, Duration, Format, Passing Score
- Who Earns This Credential - and Why
- AI-102 vs. Its Successor at a Glance
- Mapping a Prep Timeline to the Domains
- Frequently Asked Questions
- Exam AI-102 and its Azure AI Engineer Associate credential retired June 30, 2026, at 11:59 PM Central.
- The live path forward is Exam AI-103, awarding Microsoft Certified: Azure AI Apps and Agents Developer Associate.
- Generative AI and agentic solutions is the heaviest domain at 30-35% of the exam.
- The exam costs $165 USD at standard US pricing, runs 120 minutes, and requires a 700 on a 100-1000 scale.
What "AI-102" Actually Refers To Today
If you searched "what does AI-102 mean," you probably landed on conflicting information. That's because the exam code has moved. AI-102, officially Designing and Implementing a Microsoft Azure AI Solution, was the Microsoft exam that awarded the Azure AI Engineer Associate credential. It is owned by Microsoft and delivered through Pearson VUE, either at a physical test center or via online proctoring. For years, it was the go-to Microsoft certification for engineers building Azure Cognitive Services, Azure OpenAI, and search-based AI solutions.
That exam and its associated credential retired on June 30, 2026, at 11:59 PM Central, along with the online renewal assessment tied to it. It can no longer be scheduled, taken, or renewed. The name lives on mostly as search shorthand, and this site exists to help people navigate exactly that transition - see our full AI-102 Certification overview for the complete picture of how the old and new paths relate.
The Retirement That Changed Everything
Microsoft retires and replaces associate-level exams periodically as the underlying technology shifts, and AI-102 is a clean example. The retirement wasn't just a rename - it reflects a real shift in what Azure AI engineering looks like day to day. Where the old exam emphasized configuring individual Cognitive Services resources, the new exam centers on building agentic applications with Microsoft Foundry, Foundry Agent Service, and the Microsoft Agent Framework.
If you certified under AI-102 before the June 30, 2026 cutoff, the credential stays in the Active Certifications section of your Microsoft Learn transcript until its printed expiration date. There's no renewal assessment waiting for you after that date, though - once it lapses, the only way forward is earning the current credential from scratch. Anyone weighing timing questions should read our AI-102 Exam Dates 2026 guide before deciding whether to attempt anything under the old code at all.
Key Takeaway
Don't chase the old exam code. The commercially and professionally correct move is to prepare for AI-103 content now, since that's the only version of this credential you can actually earn or renew going forward.
What the Current Exam Actually Tests
Exam AI-103, Developing AI Apps and Agents on Azure, assumes no formal prerequisites, but Microsoft's stated audience profile expects app development experience in Python along with familiarity with general AI, generative AI concepts, and Azure services. If you're wondering whether you're ready to sit for it, our AI-102 Requirements 2026 article breaks down what "ready" realistically means in practice.
The skills measured are current as of April 16, 2026. Most questions cover generally available features, though commonly used preview capabilities can show up too. Candidates are expected to have working knowledge of a specific technology stack:
- Microsoft Foundry and Foundry Agent Service
- Microsoft Agent Framework
- Azure OpenAI in Foundry Models
- Azure AI Search
- Azure Content Understanding in Foundry Tools
- Azure Document Intelligence in Foundry Tools
- Azure Speech and Azure Translator
That list alone tells you this isn't a rebadged version of the old exam - it's built around orchestrating agents and models, not just calling individual APIs.
Domain-by-Domain Breakdown
The exam is organized into five weighted domains. For the full walkthrough of each, see the AI-102 Exam Domains 2026 guide - here's the shape of what you're accountable for.
Domain 1: Plan and manage an Azure AI solution (25-30%)
Covers provisioning and securing Azure AI resources, cost and responsible AI planning, and monitoring solutions in production.
- Resource provisioning, keys, and network security patterns
- Responsible AI and content safety considerations
Domain 2: Implement generative AI and agentic solutions (30-35%)
The heaviest band on the exam, and the one that most clearly separates this credential from its predecessor.
- RAG implementation patterns
- Multi-agent orchestration and function calling / tool schemas
- Conversation memory and agent evaluation
- Tracing, token analytics, and error analysis
- Autonomous workflows with approval controls
Domain 3: Implement computer vision solutions (10-15%)
Image analysis, object detection, and vision-model integration scenarios.
- Choosing between prebuilt and custom vision models
Domain 4: Implement text analysis solutions (10-15%)
Language understanding tasks such as sentiment, key phrase extraction, and translation workflows.
- Azure Translator and language service integration
Domain 5: Implement information extraction solutions (10-15%)
Structured extraction from documents and unstructured content.
- Azure Document Intelligence and Content Understanding scenarios
Notice how Domain 2 alone carries nearly as much weight as Domains 3 and 4 combined. Any prep plan that treats all five domains equally is misallocating study time - a mistake covered in more depth in How Hard Is the AI-102 Exam?
Exam Mechanics: Fee, Duration, Format, Passing Score
The logistics are straightforward but worth confirming before you register, since pricing and scheduling details shift by region and by testing modality.
| Item | Detail |
|---|---|
| Delivery | Pearson VUE, test center or online proctored |
| Standard US price | $165 USD (varies by country/region) |
| Duration | 120 minutes, may include interactive components |
| Passing score | 700 on a 100-1000 scale |
| Item count / pass rate | Not published by Microsoft |
| Retake wait (first fail) | 24 hours, longer waits for subsequent attempts |
| Renewal | Annual, free via unproctored Microsoft Learn assessment |
For a deeper look at total out-of-pocket cost including retakes and training, check AI-102 Certification Cost 2026. And if the 700 threshold has you wondering how scaled scoring actually works, AI-102 Passing Score 2026 walks through it without inventing numbers Microsoft hasn't published.
Who Earns This Credential - and Why
The profile Microsoft describes fits developers who already write application code - Python experience specifically - and want to formalize skills in building generative AI features into real products: chat interfaces backed by RAG, multi-agent systems that call tools and hand off tasks, and pipelines that extract structured data from documents or media. It's a hands-on, builder-oriented credential rather than a strategy or governance one.
Because the exam leans so heavily on agentic orchestration, teams shipping copilots, internal automation agents, and document-processing pipelines on Azure are the most natural fit for people holding this certification. If you're trying to connect the dots between the skills tested and actual job titles, AI-102 Jobs maps out where this credential tends to show up in postings, and Is the AI-102 Certification Worth It? weighs the tradeoffs honestly rather than promising a specific salary bump.
AI-102 vs. Its Successor at a Glance
| AI-102 (retired) | AI-103 (current) | |
|---|---|---|
| Exam name | Designing and Implementing a Microsoft Azure AI Solution | Developing AI Apps and Agents on Azure |
| Credential | Azure AI Engineer Associate | Azure AI Apps and Agents Developer Associate |
| Status | Retired June 30, 2026, 11:59 PM Central | Live, currently earnable |
| Associated training | AI-102T00 | AI-103T00-A (four-day course) |
| Core emphasis | Individual Cognitive Services integration | Agentic solutions, Foundry, RAG, orchestration |
This table is the fastest way to explain to a manager or teammate why "AI-102" in a job description, LinkedIn post, or training catalog may already be outdated language pointing at current-generation content.
Mapping a Prep Timeline to the Domains
Generic study advice - spaced repetition, timed practice blocks, flashcards - only matters if it's pointed at the right material in the right order. Given the domain weights, Generative AI and agentic solutions deserves the most calendar time, followed by Plan and manage an Azure AI solution, with the three smaller domains rotated through in shorter blocks.
Foundation and Domain 1
- Provision and secure Azure AI resources hands-on
- Review responsible AI and monitoring practices
Domain 2, the heaviest band
- Build a RAG pipeline and a multi-agent workflow end to end
- Practice function calling, tool schemas, and tracing/token analytics
Vision and text analysis
- Run through computer vision and text analysis scenarios
- Compare prebuilt vs. custom model choices
Information extraction and review
- Practice Document Intelligence and Content Understanding tasks
- Full timed practice run under exam-length conditions
For a more detailed week-by-week plan with resource recommendations, the AI-102 Study Guide 2026 goes further than the outline above, and pairing it with timed practice questions on our practice test platform is the fastest way to find weak spots before exam day. You can also use the full practice test library to simulate the 120-minute, interactive-question format described in the exam mechanics above.
Key Takeaway
Spend disproportionate time on Domain 2 - it's worth nearly a third of the exam and covers the newest, least-familiar material for anyone coming from the old AI-102 exam.
Frequently Asked Questions
No. Exam AI-102 and the Azure AI Engineer Associate credential retired on June 30, 2026, at 11:59 PM Central and cannot be scheduled or renewed. The current path is Exam AI-103.
Their credential remains in the Active Certifications section of their Microsoft Learn transcript until its printed expiration date, but there is no renewal assessment available after retirement.
Exam AI-103, Developing AI Apps and Agents on Azure, awards Microsoft Certified: Azure AI Apps and Agents Developer Associate.
There are no formal prerequisites, but Microsoft's audience profile assumes Python app development experience and familiarity with Azure and generative AI concepts.
The exam is 120 minutes, proctored, and may include interactive components. A passing score is 700 on a 100-1000 scale, and Microsoft does not publish a fixed item count.