Updated 1 Nov 2026 · 6 min read
On this page
What the AWS Certified AI Practitioner (AIF-C01) is and who it is for
AWS Certified AI Practitioner is AWS's foundational certification for artificial intelligence. AWS says it validates "foundational knowledge of artificial intelligence (AI), machine learning (ML), and generative AI concepts and use cases for AWS Services" and calls it a good starting point for anyone exploring AI and ML on AWS.
The exam guide is clear about the level. The target candidate has up to six months of exposure to AI and ML on AWS and "uses but does not necessarily build AI/ML solutions on AWS." Coding models, feature engineering, hyperparameter tuning and building ML pipelines are explicitly out of scope. You do need familiarity with core services such as Amazon EC2, Amazon S3, AWS Lambda, Amazon Bedrock and Amazon SageMaker AI, plus the shared responsibility model, IAM and AWS pricing models.
AWS published version 1.1 of the exam guide on 30 April 2026, and says guide updates appear about a month before they reach the exam. Version 1.1 added new objectives on token-based pricing, context engineering and foundational agentic AI concepts such as the Model Context Protocol, and brought services such as Amazon Bedrock AgentCore, Kiro and Strands Agents into scope. Older study material will have gaps.
AIF-C01 at a glance
| Fact | Detail |
|---|---|
| Exam code | AIF-C01 |
| Questions | 65 (50 scored, 15 unscored) |
| Time allowed | 90 minutes |
| Passing score | 700 on a scaled score of 100 to 1,000 |
| Exam fee | US$100 |
| Languages | Arabic, English, French (France), German, Italian, Japanese, Korean, Portuguese (Brazil), Spanish (Latin America), Spanish (Spain), Simplified Chinese and Traditional Chinese |
| Where you sit it | Pearson VUE test centre or online proctored |
| Certification valid for | 3 years |
AWS says the Italian and German versions will be retired after 15 October 2026, so sit the exam in either language by that date or pick another language.
What is on the exam
The guide splits the scored content into five domains. More than half the marks sit in the two generative AI domains.
Fundamentals of AI and ML (20%). The vocabulary and the lifecycle: how AI, ML, deep learning, GenAI and agentic AI differ, the learning types, batch versus real-time versus asynchronous inference, and when AI is the wrong answer altogether. It also covers which managed service does what (Transcribe, Translate, Comprehend, Lex, Polly) and metrics such as precision, recall and F1 score alongside business metrics like ROI.
Fundamentals of GenAI (24%). Tokens, chunking, embeddings, transformer, multi-modal and diffusion models, and the foundation model lifecycle. Expect questions on hallucinations and nondeterminism, token-based pricing, and the AWS services for building GenAI applications, from Amazon Bedrock and SageMaker JumpStart to Bedrock AgentCore. Agentic concepts such as multi-agent patterns, MCP, memory and tool use now sit here.
Applications of Foundation Models (28%). The biggest domain. You choose a model on cost, modality and latency, explain what temperature does, and know when to use RAG with Amazon Bedrock Knowledge Bases, fine-tuning, in-context learning or distillation, and what each costs. It covers prompt techniques such as few-shot and chain-of-thought, prompt risks such as jailbreaking, and evaluation with ROUGE, BLEU, BERTScore and LLM-as-a-judge.
Guidelines for Responsible AI (14%). Bias, fairness, robustness and safety, the legal risks of GenAI such as intellectual property claims, and tools like Amazon Bedrock Guardrails and SageMaker Model Cards. You also need to tell transparent, explainable models from opaque ones.
Security, Compliance, and Governance for AI Solutions (14%). Securing AI systems with IAM, encryption, Amazon Macie and AWS PrivateLink, spotting prompt injection and data leakage, grounding outputs to cut hallucinations, and governance services such as AWS Config, AWS CloudTrail and AWS Artifact.
Why people fail it
The usual mistake is treating AIF-C01 as light because it is foundational. It is not a hands-on test, but it expects you to pick the right AWS service for a business need, and many options sound plausible. A scenario that turns call recordings into text, detects sentiment and replies by voice touches Transcribe, Comprehend and Polly at once.
The second trap is the customisation ladder. Candidates know what prompt engineering, RAG, fine-tuning and pre-training are, but not which one a scenario calls for or what each costs. Read each scenario for clues about cost, effort and how current the data must be.
Third, engineers overthink. The candidate uses rather than builds, so an answer that trains a custom model is often wrong when a managed service or a model through Amazon Bedrock would do.
Finally, the formats cost marks. Multiple response, ordering and matching questions only score if every part is right. There is no penalty for guessing, so never leave a question blank.
A study plan that fits the exam
Six weeks suits most people who already know basic AWS. The weeks follow the domain weights.
- Week 1: Fundamentals of AI and ML. Terms, learning and inference types, and the managed AI services. Try the free AIF-C01 practice questions cold to see where you stand.
- Week 2: Fundamentals of GenAI. Tokens, embeddings, the model lifecycle, limitations, token pricing and the agentic AI concepts added in version 1.1.
- Week 3: GenAI on AWS and foundation model design. Amazon Bedrock, SageMaker JumpStart, AgentCore, model selection, inference parameters, RAG and vector stores.
- Week 4: Applications of Foundation Models. Prompt techniques and risks, fine-tuning, the cost of each customisation approach, and the evaluation metrics.
- Week 5: Responsible AI plus security, compliance and governance. Guardrails, Model Cards, bias and variance, IAM, encryption and the governance services.
- Week 6: Timed practice. Work through the 571-question AIF-C01 practice pack in timed sets, read the explanation for every miss, and revisit your weakest domain before booking.
On exam day
You sit AIF-C01 at a Pearson VUE test centre or online with a live proctor. At a test centre, arrive 15 to 30 minutes early with identification whose first and last names match your AWS Certification Account. Online, you can launch up to 30 minutes early, but if you are more than 15 minutes late, or your computer fails the system test, you lose the fee.
AWS does not schedule breaks. Online, you cannot leave the camera's view at all. At a test centre you can take an unscheduled break, but the clock keeps running. With 65 questions in 90 minutes you have roughly 80 seconds each, so flag hard questions and come back. Fifteen questions are unscored and not identified, so do not panic over an odd one. If English is not your first language and you test in English, you can request the ESL +30 accommodation for 30 extra minutes before you register. Results reach your AWS Certification Account within five business days.
Frequently asked questions
What happens if I fail AIF-C01?
You must wait 14 calendar days before retaking it. AWS sets no limit on attempts, but you pay the full fee each time. If your score report includes section-level feedback, use it to find weak domains, though AWS advises caution when reading it.
Do I need to know how to code?
No. The exam guide lists coding AI and ML models, data engineering and hyperparameter tuning as out of scope. You need to understand concepts, services and tradeoffs.
How do I keep the certification current?
It is valid for three years. To renew before it expires, pass the latest version of AIF-C01, the AWS Certified Machine Learning Engineer Associate exam or the AWS Certified Generative AI Developer Professional exam. Any of the three recertifies it for another three years.
Is the practice pack enough on its own?
Honestly, no. The pack is practice, not a course. It shows how AWS frames questions and where your gaps are, but pair it with the official exam guide and AWS training so you understand why each answer is right. Start with the free sample to see whether the format suits you.
When you are ready to test yourself properly, get the full AIF-C01 practice pack, with a full refund if you do not pass.
Sources
The official pages the facts in this guide were checked against.
- aws.amazon.com/certification/certified-ai-practitioner/
- docs.aws.amazon.com/aws-certification/latest/ai-practitioner-01/ai-practitioner-01.html
- docs.aws.amazon.com/aws-certification/latest/ai-practitioner-01/aif-01-revisions.html
- docs.aws.amazon.com/aws-certification/latest/ai-practitioner-01/ai-practitioner-01-domain1.html
- docs.aws.amazon.com/aws-certification/latest/ai-practitioner-01/ai-practitioner-01-domain2.html
- docs.aws.amazon.com/aws-certification/latest/ai-practitioner-01/ai-practitioner-01-domain3.html
- docs.aws.amazon.com/aws-certification/latest/ai-practitioner-01/ai-practitioner-01-domain4.html
- docs.aws.amazon.com/aws-certification/latest/ai-practitioner-01/ai-practitioner-01-domain5.html
- aws.amazon.com/certification/policies/during-testing/
- aws.amazon.com/certification/policies/before-testing/
- aws.amazon.com/certification/policies/after-testing/