What is on the AIGP exam: the 2026 blueprint, domain by domain

The AIGP exam tests 4 domains and 13 competencies, with a published item range for each. The heaviest single competency is IV.C, governing the deployment and use of an AI system, at 9 to 11 items. The lightest sit at the foundations end, including I.A, understanding what AI is and why it needs governance, at 4 to 6 items.

The blueprint you should be studying from

The current document is the IAPP AIGP Body of Knowledge and Exam Blueprint, version 2.1. The IAPP approved it on 9 September 2025 and it takes effect on 2 February 2026, superseding version 2.0.1. If you sit the exam on or after that date, version 2.1 is the map. Check the version number on anything you study from, because an outdated blueprint will have you weighting the wrong material.

The exam format

The IAPP publishes the format in its AIGP Study Guide:

"The AIGP is a three-hour exam composed of 100 items. All questions are multiple choice [...] There are no essay questions. Each correct answer is worth one point."

That is the complete published picture. Three hours, 100 items, multiple choice, one point per correct answer. The IAPP does not publish anything about how the questions are written or structured, so any guide that tells you what the questions look like is guessing.

The four domains at a glance

Domain	Title	Items
I	Understanding the foundations of AI governance	16 to 20
II	Understanding how laws, standards and frameworks apply to AI	19 to 23
III	Understanding how to govern AI development	21 to 25
IV	Understanding how to govern AI deployment and use	21 to 25

Two things stand out. First, the exam is weighted toward doing governance, not defining it. Domains III and IV together account for 42 to 50 of the 100 items. Second, no domain is safe to skip. Even the lightest, Domain I, carries 16 to 20 items.

Domain I: foundations, 16 to 20 items

Competency	Items
I.A Understand what AI is and why it needs governance	4 to 6
I.B Establish and communicate organizational expectations for AI governance	5 to 7
I.C Establish policies and procedures to apply throughout the AI life cycle	6 to 8

Domain I asks whether you can set up a governance function, not just describe one. Only I.A is definitional, and it is one of the two smallest competencies on the exam. The other two competencies are about action. I.B asks you to establish and communicate what your organization expects of AI governance. I.C asks you to build the policies and procedures that apply across the AI life cycle. Notice the pattern in the verbs: establish, communicate, establish again. If your preparation for Domain I stops at knowing what AI is, you have covered 4 to 6 items and left 11 to 15 on the table.

Domain II: laws, standards and frameworks, 19 to 23 items

Competency	Items
II.A Understand how existing data privacy laws apply to AI	4 to 6
II.B Understand how other types of existing laws apply to AI	4 to 6
II.C Understand the main elements of AI-specific laws	6 to 8
II.D Understand the main industry standards and tools that apply to AI	3 to 5

This is the legal domain, and its structure tells you how to divide your reading. It splits into three layers. Laws that existed before AI and now apply to it, covered in two competencies: data privacy law in II.A, and every other kind of existing law in II.B. Laws written specifically for AI, in II.C, which is the largest competency in the domain at 6 to 8 items. And the non-statutory layer of industry standards and tools in II.D, which is the smallest competency on the entire exam at 3 to 5 items.

The word to hold onto here is "apply." The competencies do not ask you to recite legal texts. They ask you to understand how those texts reach an AI system. A candidate who can name a law but cannot say what it demands of an AI use case is studying the wrong skill for this domain.

II.D deserves a note despite its size. Standards and tools are a distinct competency, separate from law. Do not fold them into your legal reading and assume you have them covered.

Domain III: governing development, 21 to 25 items

Competency	Items
III.A Govern the designing and building of the AI system	6 to 8
III.B Govern the collection and use of data in training and testing the AI model and system	6 to 8
III.C Govern the release, monitoring and maintenance of the AI system	8 to 10

Domain III follows the build life cycle in order. Design and build in III.A. Data for training and testing in III.B. Release, monitoring and maintenance in III.C. Every competency starts with the same verb: govern. You are not being examined as an engineer. You are being examined as the person who must put controls around what engineers do at each stage.

Note where the weight falls. III.C, at 8 to 10 items, is the second-heaviest competency on the exam and it covers what happens after the build: release, monitoring and maintenance. Candidates tend to spend their Domain III study time on training data because it feels like the hard part. The blueprint says the post-release phase carries more items. Governance does not end when the model ships, and neither does this domain.

III.B is also worth reading precisely. It covers data in both training and testing, and it names both the model and the system. Those distinctions are in the competency title for a reason.

Domain IV: governing deployment and use, 21 to 25 items

Competency	Items
IV.A Evaluate key factors and risks relevant to the decision to deploy the AI system	6 to 8
IV.B Perform key activities to assess the AI system	5 to 7
IV.C Govern the deployment and use of the AI system	9 to 11

Domain IV is structured as a sequence of decisions. First, IV.A: evaluate whether to deploy at all, weighing the factors and risks relevant to that decision. Second, IV.B: perform the assessment activities on the system itself. Third, IV.C: govern the system once it is deployed and in use.

IV.C is the single heaviest competency anywhere on the blueprint, at 9 to 11 items. On a 100-item exam, one competency out of 13 can account for roughly a tenth of your score. Put that next to III.C at 8 to 10 items and a clear message emerges: the exam cares most about live systems. What you must be able to do, according to the performance the blueprint describes, is govern AI that is already deployed and in use, not only AI that is being planned or built.

Where the weight sits, and where it does not

Ranked by maximum item count, the top of the blueprint is IV.C at 9 to 11 and III.C at 8 to 10. The bottom is II.D at 3 to 5, with I.A, II.A and II.B all at 4 to 6. The definitional material in I.A, understanding what AI is and why it needs governance, is among the lightest content on the exam. Most self-study begins there and stays there too long. The blueprint rewards the opposite allocation.

What the IAPP does not publish

Be clear-eyed about the limits of the public record. The IAPP does not publish the registration fee, the passing score, the retake policy or the recertification requirements in the blueprint or study guide. It also does not publish anything about question style. Any resource that quotes a passing score or describes typical question structure is asserting something the certification body has not stated. Treat those claims accordingly. Credential Press is independent of the IAPP, and everything above comes from the IAPP's published blueprint and study guide, nothing more.

Where this leaves you

The blueprint is a budget, and the exam will spend it as written. Between 42 and 50 of your 100 items will come from Domains III and IV, and the two competencies covering live, released, in-use systems carry up to 21 items between them. Plan your study hours against the item ranges above rather than against what feels foundational, and check every fact you memorize against the version 2.1 document itself before exam day.