AI-103 Prerequisites and Skills Self-Check

The short version

AI-103 has no required prerequisite certification that we could find on Microsoft’s pages. Microsoft does expect experience developing apps with Python and familiarity with general AI, generative AI, and Azure services. This page separates what is required from what is recommended, and gives you a checklist.

What Microsoft states

Microsoft’s exam page and study guide say that candidates should have experience developing apps by using Python, and that they need to be familiar with the capabilities of general AI, generative AI, and Azure services. Neither page lists a required earlier certification. The AI-102 exam and its certification are retired, and you do not need them for AI-103.

Microsoft also describes collaboration: AI engineers work with business stakeholders, solution architects, data scientists, DevOps engineers, and cloud security engineers. That is a hint that the exam includes some security and delivery topics, not only model calls.

Required versus recommended

Required to book: a Microsoft account to schedule through Pearson VUE. Microsoft recommends a personal Microsoft account rather than a work or school account.

Expected by the exam content:

Python at a working level, so you can read SDK code and spot a wrong call.
Basic Azure knowledge: subscriptions, resource groups, identity and access, networking, and monitoring.
General AI and generative AI concepts, including prompts, embeddings, and retrieval-augmented generation.

Helpful but optional: Experience with Azure AI Search, document processing, speech, and image generation, and some DevOps practice such as pipelines.

Skills self-check

Rate yourself honestly. If you can say yes to most of these, you are ready to start the study plan. If you say no to many, start with the matching foundation topic.

Python and development

I can call an HTTP API or SDK from Python and handle the response and errors.
I can read and write JSON and understand what a schema is.
I can set up a virtual environment and install packages.

Azure basics

I know what a resource group, a subscription, and a region are.
I can explain managed identity and how it differs from storing a key.
I know the difference between giving a person access and giving an app access.
I have a general idea of private networking and why a team would use it.

Generative AI concepts

I can explain tokens, context windows, and what temperature changes.
I can describe what an embedding is and why vector search uses it.
I can describe retrieval-augmented generation and why it reduces made-up answers.
I know what it means for an AI agent to call a tool.

Services in the exam

I have used or read about image generation, speech to text, translation, and document OCR.
I know why a team would use keyword, semantic, hybrid, or vector search.

Our glossary can help with unfamiliar terms.

If you are missing something

New to Python: work through a beginner course, then build a script that calls a model through an SDK.
New to Azure: learn the portal basics, identity, and resource organization first. The Microsoft Learn site has free learning paths; start from the AI-103 study guide, which links the resources Microsoft recommends.
New to generative AI: read the fundamentals of prompts, embeddings, and retrieval before you touch the domains.

How long this takes depends on you. A few weeks of focused practice is common for the basics, but it varies a lot.

Do you need an earlier certification?

Not according to the pages we checked. Some people take an AI or Azure fundamentals exam first to learn the vocabulary, but that is a personal choice, not a Microsoft requirement. Whether you sit one is up to you.

More guides for AI-103

Next step

When the checklist looks good, open the study plan. To see the exam’s domains in detail, read domains explained, and to check your knowledge, try the AI-103 practice tests.

Try the AI-103 practice exams