AIF-C01 · AWS
AWS put agentic AI on the AI Practitioner exam
Version 1.1 of the AIF-C01 exam guide adds agentic AI to four objectives, swaps AUC for precision and recall, and brings Amazon Q and Kiro into the services you are expected to know.
AWS publishes a change history at the back of every exam guide, and the one for AWS Certified AI Practitioner (AIF-C01) now lists two versions: 1.0, published on 26 March 2026, and 1.1, published on 30 April 2026. Version 1.1 is the one the exam measures today.
The guide is unusually explicit about what happened, because it prints the old and the new wording of every objective it touched, side by side. Read down that table and one theme runs through it.
Agentic AI is now examinable
Four objectives gained the same two words. The list of basic terms in objective 1.1.1 used to end at "large language models"; it now ends at "generative AI [GenAI], agentic AI". Objective 1.1.2, which asks you to tell AI, ML, GenAI and deep learning apart, now asks you to tell them apart from agentic AI too. And the catalogue of real-world applications in 1.2.4 gained "knowledge bases, agentic AI".
That is a genuine widening of scope, not a rewording. An exam that asked you to define a large language model now expects you to say how an agent differs from one.
Inference and metrics moved too
Two smaller changes are worth knowing before you sit it:
- Inference types. Objective 1.1.3 listed batch and real-time. It now lists batch, real-time, asynchronous and serverless — the four ways you actually deploy a model on AWS.
- Model metrics. Objective 1.3.6 dropped Area Under the Curve and put precision and recall in its place, next to accuracy and F1. If you learned AUC for this exam, that effort no longer buys you a question.
The service list caught up with the product line
Objective 1.3.4 used to name four SageMaker components for the stages of an ML pipeline. It now names Amazon Bedrock, Amazon Q, Amazon Quick, Kiro and SageMaker AI — and it talks about an "AI/ML pipeline" rather than an ML one. Objective 2.3.1 pulled in Amazon Bedrock Data Automation alongside SageMaker JumpStart and PartyRock.
When a guide change reaches the exam
One line in the Revisions section is easy to miss and worth planning around: AWS states that exam guide updates are published roughly one month before they show up in the exam itself. A guide dated 30 April was therefore live in the question pool from about the end of May.
There is no version 1.2 announced. The current objectives are the ones to study, and they are the ones our AI Practitioner course is built on.
Every date we track for this and the other exams we cover is in the exam calendar.
Sources
The whole course, on your phone
Lessons you can read, audio you can listen to on the way to work, and practice that remembers what you got wrong.