Unlock your potential in AI with the AI-900: Microsoft Azure AI Fundamentals certification—the essential credential for both technical and non-technical professionals looking to enhance their understanding of Microsoft Azure and artificial intelligence. This certification covers crucial concepts such as machine learning, natural language processing, and computer vision. With the growing demand for knowledge in AI technologies, including generative AI and conversational AI, mastering the AI-900 is a significant step for anyone aiming to excel in the rapidly evolving tech landscape. Explore the exam format, key topics, and study resources to ensure you’re prepared for success in this exciting field.

What is the AI-900?

The AI-900: Microsoft Azure AI Fundamentals certification is a standard Microsoft certification for technical and non-technical developers. It covers a wide range of topics that fall under the umbrella of Azure services relating to machine learning and other AI concepts. The certification has been gaining popularity due to the increasing conversations around AI, generative AI, conversational AI, and so on.

You can find the Microsoft Learn webpage for the certification here.

The exam has a time limit of 65 minutes and consists of 30-35 questions. You need a passing score of 700 points (out of 1000) to obtain the certification. The exam costs $99, which is reasonable compared to other Microsoft exams.

For more information on the Microsoft Exam scoring system, please refer to the hyperlink here.

As do most Microsoft Certifications, there is a breakdown of the skills the exam will measure. They are as follows:

Describe Artificial Intelligence workloads and considerations (20–25%)
Describe fundamental principles of machine learning on Azure (25–30%)
Describe features of computer vision workloads on Azure (15–20%)
Describe features of Natural Language Processing (NLP) workloads on Azure (25–30%)

How should I study for the AI-900?

I need to be honest and share a secret. This exam is not technical. The AI-900: Microsoft Azure AI Fundamentals certification is designed for both technical and non-technical developers. The exam is not super challenging, but it does require preparation. The exam is not going to ask you about the details of how ChatGPT works, but it will ask you about general principles and which Azure service to use.

To prepare for the exam, I recommend using effective tools such as the Microsoft learning modules. These modules are available on the site certification webpage and provide guided articles created by Microsoft that cover everything the exam is going to cover. At the end of each section, there is a knowledge check that provides great practice problems. Make sure you do these knowledge checks and review your answers at the end of each one. This will help you understand why you got your answer right or wrong.

Another resource is the instructor-led section, of the same webpage, which offers different courses that are either live or recorded where a user can watch and learn. These are great resources, but I haven’t personally used them.

Finally, the most valuable resource that I personally have had while studying is practice problems. I cannot stress the importance of practice questions enough. If you have taken a Microsoft certification before, you know that the exams tend to have some tricky problems, and the AI-900 is no exception.

There are plenty of different resources online that have practice problems relating to these topics. Some are free, while others cost money. I’m not going to recommend a specific one, but I will say that the more problems you do, the more prepared and confident you will feel on test day, and the more successful you will be.

For more information on my fail-proof method to pass any microsoft exam in 3 steps, be sure to check out this post here.

What is on the AI-900?

In this post, we will cover two main topics: AI principles and AI workloads. These topics are not exhaustive, and there are many more topics on the exam that reach outside of these, so make sure to check out the blog post linked here, or at the bottom of this page for more information.

AI Principles

Microsoft is using and applying several overarching general ideas towards their approach to building, training, and using AI. These principles include:

Fairness
Reliability ad Safety
Privacy and Security
Inclusiveness
Transparency
Responsibility

Most of these principles are self-explanatory. However, the difference between fairness and inclusiveness can be tricky. Let’s break it down further. If the question relates to bias of an output or bias of testing or validation data, the answer is going to be Fairness. If the AI generates a sexist response based on the data or leans towards a certain income level, that violates fairness because the outputs are not fair. Inclusiveness involves making sure everybody has the ability to access and use the AI capability regardless of income or disability. For example, if a user who needs to use the AI capability is blind, the AI capability should be able to speak the output aloud instead of transcribing it. This is an example of AI being inclusive.

AI Workloads

Now, we will discuss the different services that Azure AI provides, known as AI workloads. Each workload has sub-workloads, so it’s important to understand the differences between them. If you have questions about which workload or sub-workload to use, this post will help you.

Machine Learning

The first workload we’ll cover is machine learning. It’s broken up into regression, classification, clustering and anomaly detection.

Regression

Think of it as taking past numerical data to predict a future. For example, you can take past numerical data from the stock market and try to create a prediction of what is going to happen or what the output would be.

Classification

The next workload is classification. If you see any questions regarding an option, set a bully and a pass or fail, it is going to be classification. Classifications are going to have labels specifically for the inputs to be classified or labeled to one of those. For example, you can take all this input data of students and classify them as passing or failing the class.

Clustering

The next subsection is clustering. Clustering is going to group the data into different segments, but this is going to be tricky with classification because clustering does not have specific labels. It doesn’t say “this is the past group” and “this is the failed group.” All it’s saying is “this group’s inputs are similar” and “this group’s inputs are similar.” So it’s going to group the data but it’s not going to classify them. It’s really important to understand the difference between those.

Anomaly Detection

Anomaly detection is our next subsection. This is fairly straightforward. This is exactly what happens when you your credit card where to get swiped out of the country and you get a text message from your bank that’s anomaly detection.

Computer Vision

Computer vision is used for looking at a specific image and telling you the color, resolution, location of objects in the picture, the number of objects in the picture, and a description of the picture. These are all going to fall under the computer vision workload. Note that there is something different than computer vision, and that is custom vision. I’m not going to say computer vision is out of the box, but it has some preset capabilities that are much faster to get up and running. Custom vision has similar capabilities, but you have to provide your own training data to train and test your models.

Natural Language Processing

Natural language processing handles things like language detection, language translation, text analysis for things like customer sentiment. It handles speech analysis, which takes your speech to tell asked for things like captions or and also speech synthesis, which is text to speech.

Knowledge Mining

The last AI workload is knowledge mining. Knowledge mining is used to extract information from large volumes of unstructured data to create a searchable knowledge store. Although you may not have any questions involving knowledge mining on your practice problems or on your exam, it’s still important and valuable to know what that is.

In Conclusion

The AI-900: Microsoft Azure AI Fundamentals certification is a standard Microsoft certification that covers a wide range of topics related to Azure services and machine learning. The exam is designed for both technical and non-technical developers and requires preparation. Although the exam is not super challenging, it will ask you about general principles and which Azure service to use. To prepare for the exam, I recommend using effective tools such as the Microsoft learning modules, which provide guided articles created by Microsoft that cover everything the exam is going to cover. Additionally, practice problems are an invaluable resource for studying. By following these tips, you can be well-prepared for the AI-900 exam.

Microsoft AI-900: The Complete Certification Overview