AI-102 Certification

What "AI-102" Means in 2026

For years, "AI-102" was shorthand for the exam that built Azure AI engineering skill sets: prompt design, cognitive services integration, and applied AI on Azure. Anyone searching for AI-102 certification today lands in the middle of a naming transition that matters a lot if you're planning to sit for it. Microsoft owns the credential and delivers it through Pearson VUE, either at physical test centers or via online proctoring - that delivery mechanism hasn't changed. What has changed is the exam itself and the skills it validates.

If you want the short version of where things stand right now, our What Is AI-102? explainer breaks down the naming history in plain language. This article goes deeper: what the credential covers today, how the domains are weighted, what it costs, and how the retirement of the original exam affects anyone still holding it or planning around it.

The Retirement Timeline You Need to Know

The retirement of the original exam is a hard stop, not a soft deprecation. Along with the exam and the associate credential, the renewal assessment tied to it also retired. That means:

• The exam can no longer be scheduled or taken after the retirement date.
• The credential can no longer be newly earned through that path.
• Anyone who certified before the cutoff keeps the credential in the Active Certifications section of their Microsoft transcript until its printed expiration date.
• There is no renewal path for the retired credential once it expires - holders who want to stay current need to move to the successor exam.

This is exactly the kind of detail that trips people up when they're scheduling exams months in advance. If you're mapping out testing windows, cross-check your plans against AI-102 Exam Dates 2026: Testing Windows, Deadlines & Scheduling before you commit to a study calendar built around an exam code that may no longer be bookable.

The Five Domains Behind the Exam

The current skills measured are dated April 16, 2026, and they're organized into five weighted domains. Most questions target generally available features, though commonly used preview capabilities can show up too. Here's the breakdown:

For a full walkthrough of every subskill inside each domain, see AI-102 Exam Domains 2026: Complete Guide to All 5 Content Areas. The short version: two domains - planning/management and generative AI/agentic solutions - together account for more than half the exam, so your study hours should be allocated accordingly rather than split evenly across all five.

Inside the Heaviest Domain: Generative AI and Agentic Solutions

At 30-35%, this domain deserves the largest share of your study time, and it's also the most conceptually dense. It assumes working knowledge of Microsoft Foundry, Foundry Agent Service, Microsoft Agent Framework, and Azure OpenAI in Foundry Models - tools that weren't part of the original AI-102 syllabus at all. If you're coming from an older study plan, this is the section where your material is most likely stale.

Candidates often underestimate how much of this domain is about operational rigor - tracing an agent's decision path, analyzing token consumption, and building in approval gates - rather than just getting a chatbot to respond correctly. If you want a difficulty-adjusted view of where this domain ranks against the others, How Hard Is the AI-102 Exam? Complete Difficulty Guide 2026 covers the pattern of question styles you'll encounter, including scenario-based items with interactive components.

Registration, Fees, and Exam Logistics

The mechanics of scheduling and sitting for the exam are straightforward, but a few numbers matter for planning:

• Price: $165 USD at standard US pricing; the amount varies by the country or region where the exam is proctored.
• Duration: 120 minutes, proctored, and the exam may include interactive components beyond standard multiple choice.
• Passing score: 700 on a 100-1000 scale. Microsoft does not publish a fixed item count or pass rate.
• Delivery: Pearson VUE, at a test center or via online proctoring.
• Retakes: A failed first attempt can be retaken after 24 hours, with progressively longer waits for subsequent attempts.
• Renewal: Microsoft associate certifications expire annually and renew free by passing an unproctored online assessment on Microsoft Learn.

For a complete cost breakdown including how regional pricing works, read AI-102 Certification Cost 2026: Complete Pricing Breakdown. And if you want to know exactly what 700 on that scale means for how many questions you can afford to miss, AI-102 Passing Score 2026: Exactly What You Need to Pass walks through the scoring logic in detail.

Who Actually Earns This Credential

The audience profile is developers, not data scientists in the traditional sense - people building applications and agent-based systems on Azure who need to integrate services like Azure AI Search, Azure Content Understanding in Foundry Tools, Azure Document Intelligence in Foundry Tools, Azure Speech, and Azure Translator into production software. That's a meaningfully different profile from someone doing model training or MLOps.

If you're weighing whether this is the right investment for your career stage, Is the AI-102 Certification Worth It? Complete ROI Analysis 2026 lays out the tradeoffs without inflating expectations, and AI-102 Jobs looks at where this skill set gets applied on real teams. For eligibility specifics - including what "no formal prerequisites" actually means in practice - see AI-102 Requirements 2026: Eligibility, Prerequisites & How to Qualify.

A Domain-Aware Study Timeline

Generic study advice rarely accounts for the fact that one domain here carries more weight than three others combined. A study plan built around this exam's actual weighting looks different from a flat, evenly-split schedule.

Because domain 2 alone is worth as much as the two smallest domains combined, front-loading it early and revisiting it in week 6 tends to pay off more than treating every domain equally. A more detailed week-by-week breakdown, including how to sequence practice exams, is in AI-102 Study Guide 2026: How to Pass on Your First Attempt.

Migrating From AI-102 Knowledge to the Current Exam

If you studied older AI-102 material - cognitive services fundamentals, classic prompt engineering, earlier Azure OpenAI patterns - a lot of that foundation still transfers. What's new is the agentic layer: Microsoft Foundry as the unifying platform, Foundry Agent Service for orchestration, and Microsoft Agent Framework for building multi-agent systems. The training also shifted: the AI-103T00-A four-day course replaced the earlier AI-102T00 course at the training level, reflecting how much of the curriculum is now agent-centric rather than single-model-centric.

Practically, this means your prep time should be reallocated rather than discarded. Core service knowledge (Speech, Translator, Search) still applies, but you need to layer agent orchestration, evaluation, and tracing on top. Running timed practice questions against the current domain weighting is the fastest way to find those gaps - you can start with a full-length simulation on AI-102 Exam Prep's practice test platform and see exactly where the agentic-solutions questions expose weak spots.

Before your first attempt, it's worth reviewing a condensed reference of the must-know facts across all five domains - our AI-102 Cheat Sheet 2026: One-Page Review of Must-Know Facts is built for exactly that kind of last-mile review, and pairing it with timed drills on the practice test site tends to catch gaps that flashcards alone miss.

If you're also weighing the long-term earning picture around this specialization, AI-102 Salary Guide 2026: Complete Earnings Analysis and AI-102 Training cover how employers value this skill set and where formal training fits into a self-study plan. And if you'd rather see how other candidates' outcomes trend before committing serious study hours, AI-102 Pass Rate 2026: What the Data Shows is a useful companion read alongside a round of practice questions on AI-102 Exam Prep.

Frequently Asked Questions