AWS Certified AI Practitioner
AI Practitioner is AWS testing whether you can discuss machine learning and generative AI accurately: the vocabulary, the AWS services that deliver it, and the responsible-use questions that decide whether a project survives a legal review.
Ninety minutes, 65 questions, no prerequisite. It sits at the same tier as Cloud Practitioner and shares its ambition.
Exam facts
- Exam code
- AIF-C01
- Level
- Foundational
- Field
- Data & AI
- Duration
- 90 minutes
- Questions
- 65 questions (50 scored, 15 unscored); multiple choice, multiple response and further item types
- Passing score
- 700 of 1,000 (scaled)
- Languages
- English, German, French, Spanish, Italian, Portuguese, Japanese, Korean, Chinese, ArabicAWS retires the German and Italian editions after 15 October 2026. The badges group regional variants: Spanish (Spain and Latin America) and Chinese (Simplified and Traditional) count as one language here, so the vendor lists twelve exam editions where this page shows ten.
- Price
- 100 USD (read 12 September 2026)
- Validity
- 3 years, renewable
- Exam delivery
- Pearson VUE (test centre or online proctored)
Prerequisites: None. AWS recommends about six months of exposure to AI or ML solutions on AWS — from business, sales or project roles as much as from IT.
Renewal: Pass the current version again (50 % voucher in your AWS certification account) or pass a higher-tier AWS exam
Source: exam page at AWS · Price: AWS quotes exam prices in US dollars; Pearson VUE charges the local equivalent at booking using the rate AWS sets
How to prepare
1. The vendor's free learning path
AWS publishes the exam objectives and a learning path free of charge — the authoritative source for scope and weighting. Open the learning path (opens in a new tab)
2. Online courses
Courses and practice questions from the directory that target exactly this exam — details, price and the provider link (affiliate) are on the course page.
3. A practice test in the exam format
MeasureUp sells a practice test for AIF-C01 with questions in the exam format, an explanation for every answer and a timed mode; the price shows in your currency on the shop page. Affiliate link.
Practice test for AIF-C01 at MeasureUp (opens in a new tab)Book the exam
Delivered by Pearson VUE (test centre or online proctored). Schedule with AWS (opens in a new tab)
Affiliate disclosure: the course and practice-test links above are affiliate links — buying through them may earn us a commission at no extra cost to you. The vendor's learning path and booking links carry no commission.
A vocabulary exam in a field with too much vocabulary
The genuine value here is unglamorous. Terms like inference, embedding, fine-tuning, hallucination and prompt injection get used loosely in meetings, and somebody who uses them precisely is immediately more useful than somebody who does not. The exam also spends real time on responsible use, bias and governance, which is the part practitioners skip and the part that stops deployments.
Five domains, and the practical one is the largest
- Fundamentals of AI and machine learning (20 %): what a model is, what training does, and the difference between the kinds of learning.
- Fundamentals of generative AI (24 %): tokens, prompts, foundation models, and where they go wrong.
- Applications of foundation models (28 %): the largest share, and the one that assumes you have watched a real use case be scoped, costed and evaluated.
- Guidelines for responsible AI (14 %): bias, transparency, and who is accountable when the output is wrong.
- Security, compliance and governance (14 %): the domain that decides whether a project survives a legal review.
Together the two governance domains are worth more than a quarter of the paper, and they are where technical candidates lose marks, because the questions are about process and accountability rather than about models. The applications domain rewards having seen a real project: which model class fits a task, how a result is evaluated, and what the running cost depends on. None of it requires mathematics. All of it requires having listened carefully in the meetings where these words were used badly.
| Aspect | AI Practitioner (AIF-C01) | Cloud Practitioner (CLF-C02) |
|---|---|---|
| Time and questions | 90 minutes, 65 questions | 90 minutes, 65 questions |
| Pass mark | 700 of 1,000 | 700 of 1,000 |
| Fee | 100 USD | 100 USD |
| Validity | Three years | Three years |
| German and Italian editions end | After 15 October 2026 | After 31 December 2026 |
| Renewal | Resit or a higher exam | Resit, a higher exam, or a free game |
Two language editions are being withdrawn
AWS has announced that the German and Italian editions retire after 15 October 2026. If either is your language, that is a booking deadline rather than a detail, and it is the kind of change that appears on a vendor page without appearing in any course. After the cut-off the exam remains available in English and the other supported languages, and non-native English speakers can request additional time, which must be arranged before booking rather than on the day.
Who should sit it, and who should look one tier up
AWS recommends about six months of exposure to AI or machine-learning work on its platform, and it says explicitly that business, sales and project roles count as much as IT. Take that at face value: this is the certificate for the product manager, the procurement lead, the compliance officer and the salesperson who have to talk about these systems accurately and be trusted to. Engineers who build models or pipelines should treat it as optional; the associate data exams test what they do, and this one tests whether they can explain it. Three years of validity and a free renewal through any higher AWS exam keep the downside small if the field moves underneath it, which it will.
What a question looks like
Written by us in the exam's style. It is not a real question from any question bank, and we do not publish those.
A bank wants an assistant that answers staff questions using its own internal policy documents, which are revised monthly. Accuracy against the current policy matters more than fluency. Which approach best fits the requirement?
Fine-tuning sounds like the sophisticated answer and is the wrong one, because monthly revisions would mean retraining on a schedule nobody will keep. The question rewards understanding that grounding a model in retrieved documents separates the knowledge from the model, which is a conceptual distinction rather than a technical one and is exactly what this tier is testing.
What it costs to get and to keep
| Item | Amount | Note |
|---|---|---|
| Exam fee | 100 USD | quoted in US dollars and converted by Pearson VUE at booking (read 12 September 2026) |
| Retake | 100 USD | payable in full after the waiting period (read 12 September 2026) |
| Recertification every 3 years | not published | AWS does not publish a fixed recertification price for this exam; passing any higher AWS exam renews it, which is the cheaper route if you intend to continue (read 12 September 2026) |
How much preparation, from where you are
- You already use AI tools at work but did not build them
- You will know more than you think and use the words loosely. The work is tightening definitions and learning which AWS service corresponds to each idea, plus the governance material you have probably never read.
- You are a developer or data practitioner
- The concepts will be familiar and the AWS product names will not. Study the service catalogue and the responsible-use section, and consider whether an associate-level data exam would say more about you.
- You sell, buy or govern AI systems
- This is the sharpest fit on the site for that role. The exam's weakest area for engineers, governance and bias, is the part of your job that carries actual risk.
What passing this does not prove
- Any ability to train, tune or evaluate a model.
- Whether you can write code that calls these services.
- Whether an AI system you specified would survive contact with real users or real data.
- Anything that is not on AWS. The concepts travel; the product names do not.
Against the alternatives
- AI-901 — Microsoft Certified: Azure AI Fundamentals
- The direct Microsoft counterpart, and the choice should follow your employer's platform rather than the syllabus. Holding both says less than most people assume.
- CLF-C02 — AWS Certified Cloud Practitioner
- The same tier for cloud rather than AI. If you are unsure which gap you have, the honest answer is usually cloud first, because the AI services sit on top of it.
- DEA-C01 — AWS Certified Data Engineer – Associate
- A genuine step up, and a better destination for anyone who will touch the data these systems learn from. It certifies work rather than vocabulary.
AIF-C01 — quick answers
Will this get me an AI job?
By itself, no, and it was never meant to. It signals that you can be trusted to discuss these systems accurately, which matters in product, procurement, compliance and sales roles. For engineering work employers look for something you have built and can explain.
Do I need machine-learning maths?
No. That is a real distinction between this tier and the associate data exams. You need to know what a model does, what it costs, how it fails and who is accountable, none of which requires linear algebra.
Is it too soon to certify in a field moving this fast?
The vocabulary is more stable than the products, which is the strongest argument for the foundational tier specifically. Three years of validity and a free renewal through a higher exam limit the downside if the field moves underneath it.
Cloud Practitioner or this one first?
Whichever meeting you understand less of. The two sit at the same tier with the same time, price and pass mark. If your gap is the platform, take the cloud paper; if it is the models, take this. A second foundational badge adds little to the first, because each already implies the other's level.
How many languages is it really in?
AWS lists twelve exam editions and this page shows ten, because the badges group Spanish for Spain with Spanish for Latin America and the two Chinese scripts together. Arabic, English, French, German, Italian, Japanese, Korean and Portuguese make up the rest, with German and Italian ending after 15 October 2026.
What comes next
Concepts this exam draws on
Glossary entries with the reason each one matters for AIF-C01.
- GenAI
The exam's core domain: what generative models do, where they fail and how AWS packages them.
- LLMs
Foundation-model concepts (tokens, context, fine-tuning versus prompting) are a fifth of the questions.
- Prompt Engineering
AIF-C01 tests prompt techniques and their risks as applied skills, not as theory.
- RAG
Bedrock knowledge bases are the exam's example of grounding a model in your own data.
- Responsible AI
A named domain: bias, explainability, guardrails and the governance AWS expects you to recognise.
Last reviewed 12 September 2026 · Getting Digital
