How to Pass the AWS AI Practitioner (AIF-C01): A 3-Week Study Plan

About 27 hours over three weeks, weighted to the five domains in the official exam guide. AWS lists the AI Practitioner at 100 USD, 65 questions in 90 minutes, and a passing score of 700 on a 100 to 1,000 scale. Only 50 of the 65 questions count; 15 are unscored items AWS is testing, and you will not know which ones.

This is a foundational exam. Nobody asks you to write code or size a GPU cluster. The target candidate has up to six months of exposure to AI and ML on AWS, and the questions are about picking the right service, the right model-customization approach, and the right guardrail for a described situation. That is a vocabulary and judgment exam, and three weeks is enough if you spend the hours where the blueprint puts them.

If you are new to AWS itself, add a few days up front. AWS recommends Cloud Practitioner Essentials or Technical Essentials before this exam, and the service names alone will slow you down otherwise.

Before week one: two things

Download the official exam guide PDF from the AWS certification page. It has the full task-statement list per domain and the in-scope service list, which is the closest thing to a syllabus you will get.

Know the weights, because this plan follows them.

Domain	Weight	Plan hours
Fundamentals of AI and ML	20%	5
Fundamentals of Generative AI	24%	6
Applications of Foundation Models	28%	7
Guidelines for Responsible AI	14%	3.5
Security, Compliance, and Governance for AI Solutions	14%	3.5

That is 25 hours of study. The last 2 hours are practice questions and a timed run, which puts the total at about 27.

Week 1: the two fundamentals domains (about 11 hours)

Domains 1 and 2 are 44% of the exam and they are mostly definitions, so get them out of the way while your motivation is highest.

Follow the exam prep plan on AWS Skill Builder. AWS points every candidate to it from the exam page, and it walks the blueprint in order.
Build a one-page glossary in your own words: supervised versus unsupervised versus reinforcement learning, inference, embeddings, tokens, context window, temperature, hallucination, RAG, fine-tuning. Write the AWS service next to each concept where one exists.
Learn the ML lifecycle as AWS describes it, from data collection through monitoring, and the SageMaker feature that matches each stage.

By the end of the week you should be able to read a short scenario and say which category of learning it is and whether generative AI is even the right tool for it. A surprising number of questions are really asking that second thing.

Week 2: foundation models and responsible AI (about 10.5 hours)

Domain 3 is the biggest single slice at 28%, and it is where the exam gets specific. Spend 7 hours here.

Learn the customization ladder in cost order: prompt engineering, then RAG, then fine-tuning, then training from scratch. The exam loves a scenario where the cheapest adequate option is the right answer.
Know Amazon Bedrock at the level of what it does and what its guardrail, knowledge base and agent features are for. You do not need console clicks, you need to know which feature solves which described problem.
Know the inference parameters and what moving each one does to output.
Read the prompt engineering section of the guide twice. Zero-shot, few-shot, chain of thought, and prompt injection show up as answer choices.

Then 3.5 hours on Responsible AI. Bias, fairness, explainability, transparency, and which AWS tools (model cards, Clarify, guardrails) map to each. This domain is short and the questions are direct.

Week 3: security, then practice (about 6 hours)

Domain 5 is 14% and it overlaps with anything you already know about AWS security. IAM, encryption at rest and in transit, VPC endpoints, CloudTrail, and the shared responsibility model, applied to AI workloads. Add data governance: lineage, retention, and residency.

Then do practice. AWS publishes an Official Practice Question Set and an Official Pretest through Skill Builder, and the exam page tells you to use both. Take a full timed set. 90 minutes for 65 questions is about 83 seconds each, which is plenty, but the ordering, matching and case-study formats take longer than plain multiple choice, so practice those specifically.

Review every miss by domain. If any domain is under 70%, that is where the last few hours go.

Exam day

The exam is delivered by Pearson VUE, at a test center or online proctored. Online, plan for a 30-minute early check-in: system test, government photo ID matching your registration name, and a room scan. Single monitor, clear desk, no headphones.

Formats: multiple choice, multiple response, ordering, matching, and case study. There is no penalty for a wrong answer, so answer everything.

If you fail, AWS makes you wait 14 calendar days before trying again, there is no cap on attempts, and each attempt is the full 100 USD. The certification is valid for 3 years.

One 2026 note from AWS: the Italian and German versions of AIF-C01 retire after October 15, 2026. Ten other languages, including English, Japanese and both Spanish variants, continue.

The habit that passes this exam

Every practice question you miss, ask which domain it belongs to and which word in the stem you skipped. Most AIF-C01 misses are not knowledge gaps. They are a missed qualifier like lowest cost, least operational effort, or without retraining, which changes the correct answer. Train yourself to underline it before you look at the options, and the 700 comes.

Drill it week by week

Free AIF-C01 practice questions, weighted to the official blueprint, with a plain-English explanation on every one.

Free practice questions →

Keep reading: AIF-C01 study guide · Is the AI Practitioner worth it? · AWS AI Practitioner vs Azure AI Fundamentals · All AI & cloud exam prep

HOW TO // AI is not affiliated with or endorsed by AWS. AWS Certified AI Practitioner, AIF-C01 and AWS are trademarks of Amazon Web Services, Inc. or its affiliates; we reference them descriptively. All content is original.