What is the AWS AI Practitioner exam, and is it worth taking?
by certbrain-system · October 19, 2026
Welcome to the CertBrain blog. CertBrain exists to help you pass one exam, the AWS Certified AI Practitioner, so this first post covers the basics: what the exam is, who it's for, and what you need to know before you book it.
Every fact about the exam below comes from AWS's official exam guide and certification page. AWS updates both from time to time, so check them before you book.
What the exam is
The AWS Certified AI Practitioner (exam code AIF-C01) is Amazon Web Services' foundational certification in artificial intelligence. According to AWS, it's for people who want to show a foundational understanding of AI concepts and AWS AI tools, with a focus on practical business uses of AI.
The exam checks that you can:
- describe AI, machine learning and generative AI concepts, both in general and on AWS
- recognise when AI is the right tool for a business problem
- pick the right type of AI technology for a given use case
- use AI responsibly
Who it's for
AWS describes the ideal candidate as someone with up to six months of exposure to AI and machine learning on AWS, who uses AI solutions but doesn't necessarily build them. The roles AWS lists range from cloud, development, data and IT to line-of-business roles.
What the exam doesn't expect matters just as much. AWS lists these as out of scope: coding models or algorithms, data or feature engineering, tuning models, building ML pipelines, and the maths and statistics behind models. If you've been put off AI certifications because you don't write code, this one is designed for you.
AWS does expect basic familiarity with core AWS services (such as Amazon S3, Amazon EC2, AWS Lambda, Amazon Bedrock and Amazon SageMaker AI), the shared responsibility model for security, IAM for controlling access, and how AWS prices its services. If you're new to AWS altogether, AWS suggests starting with its introductory courses, such as AWS Cloud Practitioner Essentials.
Why take it
- It gives you a shared vocabulary. Many people now work alongside AI projects without building them: product managers, analysts, salespeople, consultants. The exam covers the terms and trade-offs those conversations turn on, such as foundation models, fine-tuning, retrieval-augmented generation, hallucinations, bias and responsible AI.
- It's tied to the tools many teams already use. Rather than staying abstract, the exam asks which AWS service fits which job, which is useful if your company already runs on AWS.
- It's a manageable first step. It's a foundational exam with no coding, and it gives you a structured reason to learn the subject properly rather than picking it up piecemeal.
- It lasts. The certification is valid for three years.
The exam at a glance
| Cost | 100 USD (additional costs may apply) |
| Time | 90 minutes |
| Questions | 65: 50 scored and 15 unscored |
| Passing score | 700, on a scale of 100 to 1,000 |
| Where | A Pearson VUE test centre, or online with a remote proctor |
| Valid for | 3 years |
The 15 unscored questions are ones AWS is trying out for future exams. They aren't marked, so you can't tell which they are. Treat every question as if it counts.
How it's scored
Three rules from the exam guide are worth knowing before you start:
- There's no penalty for guessing. An unanswered question is marked wrong, so never leave one blank.
- You only need to pass overall. The scoring is compensatory: a strong score in one domain can make up for a weaker one. You don't need to pass each section separately.
- Your score report shows strengths and weaknesses by section, but AWS advises caution in reading too much into those section-level results.
The question types
Most questions are multiple choice: one correct answer and three wrong ones. The exam can also include:
- Multiple response: two or more correct answers out of five or more options. You must pick all the correct ones to get the mark.
- Ordering: put three to five steps in the right order.
- Matching: match three to seven prompts with their answers. Every pair must be right.
Most questions are short scenarios: a business wants to do something, and you choose the approach or AWS service that fits best. Knowing definitions helps, but recognising which concept a scenario is describing helps more.
The five domains
The exam guide divides the content into five domains, each worth a fixed share of the scored questions:
| Domain | Share of scored questions |
|---|---|
| 1. Fundamentals of AI and ML | 20% |
| 2. Fundamentals of Generative AI | 24% |
| 3. Applications of Foundation Models | 28% |
| 4. Guidelines for Responsible AI | 14% |
| 5. Security, Compliance and Governance for AI Solutions | 14% |
In short:
- Domain 1 covers the basics: what AI, machine learning and deep learning are, the types of learning, when AI is (and isn't) the right tool, and the stages of a machine learning project.
- Domain 2 covers generative AI: how foundation models and large language models work, what they're good and bad at, and the AWS services for building with them.
- Domain 3, the largest, covers putting foundation models to work: choosing a model, writing prompts, retrieval-augmented generation, fine-tuning, and judging how well a model performs.
- Domain 4 covers responsible AI: bias, fairness, transparency and explaining how a model reaches its answers.
- Domain 5 covers securing AI systems, meeting compliance requirements and governing how AI is used.
Domains 2 and 3 together make up more than half the exam, so generative AI deserves the most study time.
How to prepare
- Find your gaps first. Before reading everything from the start, find out what you already know. CertBrain's readiness diagnostic is a short quiz across all five domains that shows where you're strong and where to focus. It's free and needs no account.
- Study by domain, weighted by the exam. Spend your time roughly in proportion to the weightings above, with extra on any domain the diagnostic flags.
- Practise on scenarios, not just definitions. The exam tests whether you can apply an idea to a situation, so practise with questions that do the same.
- Review on a schedule. Reading a lesson once isn't enough to remember it weeks later. Spaced repetition brings each idea back just before you'd forget it, which is why CertBrain builds every course around it.
- Read AWS's own material. The exam guide is short and free, and it lists exactly which AWS services are in scope.
How CertBrain helps
CertBrain has one course for each exam domain: 70 short lessons in total, each followed by scenario questions in the exam's style and flashcards on the key terms. Everything you study comes back for review on a spaced-repetition schedule.
All of it is original, written from AWS's public documentation, and every question records the AWS documentation page it's based on. Where AWS has retired a service, we say so, and a retired service is never the correct answer.
Ready to see where you stand? Take the free diagnostic.
CertBrain is an independent study resource. It isn't affiliated with or endorsed by Amazon Web Services. AWS and AWS Certified AI Practitioner are trademarks of Amazon.com, Inc. or its affiliates.