Azure AI Engineer Certification Guide

If you searched for an AI-102 exam guide in late 2026, the most important thing to know comes first: Microsoft has retired the Azure AI Engineer Associate certification and its AI-102 exam. The official certification page on Microsoft Learn now carries the notice "This certification and the renewal assessment are retired", which means new candidates can no longer book the exam and existing holders can no longer renew. Secondary sources place the retirement at the end of June 2026, with a successor credential focused on AI apps and agents expected in its place — more on that below.

That does not make this page pointless. Thousands of developers were mid-preparation when the change landed, many employers still list "Azure AI Engineer" in job adverts, and almost everything AI-102 measured — deploying Azure AI services, building generative AI solutions, computer vision, natural language processing — still matters in day-to-day Azure work. This guide explains what the certification was, exactly where it stands now, what Microsoft's shake-up means for you, and how to redirect your study effort so none of it is wasted.

The status of AI-102 in 2026, in plain terms

Microsoft is carrying out its largest certification overhaul since 2019, retiring roughly a dozen role-based certifications between mid and late 2026 and steering the portfolio towards AI-centred credentials. AI-102 is one of the casualties, alongside the developer-focused AZ-204 and the data-science exam DP-100. Here is what that means in practice:

• You cannot book or sit AI-102 any more. The certification and its renewal assessment are officially retired on Microsoft Learn.
• If you already hold Azure AI Engineer Associate, it stays on your transcript permanently. Retired certifications are not deleted. You simply cannot renew it, so once your current 12-month validity window lapses, the credential shows as historical rather than active.
• There is no automatic migration. Reporting on the overhaul indicates there is no free upgrade, voucher, or discount that converts an AI-102 credential into its successor — you would earn the new certification from scratch.
• A successor is expected. Secondary sources name a new exam, AI-103, leading to an Azure AI App and Agent Developer Associate credential. Microsoft had not published a full Learn page for it that we could verify at the time of writing, so treat the name, format and objectives as provisional until they appear on learn.microsoft.com.

Because the successor's details are not yet official, the honest advice is to check the Microsoft Learn certification catalogue before committing to a study plan, and to treat any third-party claims about the new exam's question style or difficulty with scepticism.

What the Azure AI Engineer Associate certification measured

Understanding what AI-102 tested is still useful, both because the skills remain current and because the successor credential is expected to cover overlapping ground. The exam — formally "Designing and Implementing a Microsoft Azure AI Solution" — was a 100-minute proctored test, scored on Microsoft's standard 1–1,000 scale with a passing mark of 700. Like all Microsoft technical exams, that 700 is a scaled score, not 70% of questions correct, and Microsoft never published an official question count.

In its final published form, the exam's domains were:

• Plan and manage an Azure AI solution. Selecting the right Azure AI service for a requirement, provisioning resources, managing keys and security, monitoring usage and cost, and deploying containers where data residency demanded it.
• Implement generative AI solutions. Working with Azure OpenAI Service models, prompt engineering, grounding responses in your own data, and applying responsible-AI controls to generated content.
• Implement an agentic solution. The most recent addition, reflecting Microsoft's pivot to AI agents that orchestrate tools and act on a user's behalf rather than simply answering prompts.
• Computer vision. Image analysis, object detection and custom vision model training with Azure's vision services.
• Natural language processing. Language detection, sentiment analysis, entity recognition, translation, speech-to-text and text-to-speech, and conversational language understanding.
• Knowledge mining. Building search and document-intelligence pipelines that extract structured information from unstructured content.

The audience was software developers — people comfortable in C# or Python calling REST APIs and SDKs — rather than data scientists building models from the ground up. That distinction matters when you choose where to go next: the retired data-science counterpart, DP-100, served a different reader, and its story is covered in our Azure Data Scientist certification guide.

Who this retirement actually affects, scenario by scenario

The right move depends on where the retirement caught you. Four common situations:

You were mid-study for AI-102 and never sat it. Do not bin your notes. The service-level skills — Azure OpenAI deployments, vision and language APIs, search pipelines, responsible AI — are precisely the areas Microsoft says its new AI-era credentials emphasise. Keep the hands-on knowledge, pause spending on AI-102-specific materials, and wait for the successor's official skills outline before buying anything new.

You hold Azure AI Engineer Associate and it is still within its validity window. Your credential remains genuine and worth listing while valid. Plan for the fact that you cannot renew it: if an active AI certification matters for your role, budget time for the successor exam once Microsoft publishes it.

You are new to Azure AI and were about to start here. Start one rung lower instead. The Azure AI Fundamentals certification is active, never expires, and was refreshed in April 2026 — the old AI-900 exam has been replaced by AI-901, but the certification itself continues, now covering AI concepts and building solutions with Microsoft Foundry (the renamed Azure AI Foundry). It is the safest Azure AI investment you can make today, and ExamPractice hosts Azure AI Fundamentals practice questions you can use to check your grounding before the broader landscape settles.

You need a certification for a job advert that names "Azure AI Engineer". Employers' job templates lag certification catalogues by months or years. Be straightforward in applications: the credential is retired, you hold it (or hold the skills), and Microsoft's replacement is pending. Hiring managers dealing with Azure day to day will already know the portfolio is in flux.