Which four domains does the Standard use?
AIRS-J/1 groups its nine dimensions under four domains: Structure, Test, Engage and Perform. This page names the dimensions in each domain and explains how a domain counts in a certification result.
Judgment is how well a person decides once AI has given the answer. AIRS-J/1, the Standard that AIR APAC publishes, describes judgment in nine dimensions. Section 5 states the structure in one line: "Nine dimensions, grouped under four STEP domains."
A domain is "One of four groupings of dimensions under the STEP structure: Structure, Test, Engage, Perform" (Section 3). Each dimension receives its own score. A domain brings related dimensions together. Rule 2 of the decision rules applies to each domain (Section 12.2).
The four domains
The line under each domain name is the Standard's own (Section 5). The plain readings are AIR APAC's short explanations; the official definitions are in Section 5.
Structure
Section 5: "Framing the problem before reaching for a tool."
Dimensions: S1 Problem Framing and S2 Scope and Constraint Definition.
Structure is about the goal: what the work has to achieve. Plain reading: S1 names the real decision to be made now, as distinct from the request as it was handed over. S2 sets what is fixed, what is open and what is unknown, and what information would be enough to go ahead.
Definitions: nine dimensions page.
Test
Section 5: "Interrogating evidence, including machine-generated evidence."
Dimensions: T1 Evidence Verification, T2 Assumption and Alternative Testing and T3 Risk and Failure Anticipation.
Test is about evidence: what shows that a claim or an answer is supported, or wrong. Plain reading: T1 traces a claim to its source and checks in proportion to the consequence. T2 names the assumptions and tests them, including against the person's own view. T3 names how the decision could fail, who pays, and which harms are being accepted.
Definitions: nine dimensions page.
Engage
Section 5: "Authority, people, and being seen to act."
Dimensions: E1 Decision Rights and Escalation and E2 Stakeholder Alignment and Disclosure.
Engage is about agreement: who must agree, and who must know. Plain reading: E1 acts within the person's own authority and escalates what needs another party's approval. It does not hand to a person or a machine what cannot be handed over. E2 consults those who must be consulted and informs those who must be informed. It also discloses AI use, uncertainty and interest without being asked.
Definitions: nine dimensions page.
Perform
Section 5: "Deciding proportionately, and owning the decision."
Dimensions: P1 Proportionate Commitment Under Uncertainty and P2 Consequence Ownership and Correction.
Perform is about responsibility: being answerable for the action and what follows. Plain reading: P1 commits to a defensible position in the time available, at a cost in proportion to the evidence, and states the basis. P2 owns the outcome without passing it to the AI or the data. It also separates a poor decision from an unlucky outcome.
Definitions: nine dimensions page.
How a domain counts in a result
Each dimension is scored from 0 to 4, and Level 2, Competent, is the threshold (Section 7). The domains then count in four ways.
- A domain score is an average. Section 12.2.1 defines it as "the arithmetic mean of its dimension scores". Each dimension score is in turn the mean of its valid observations.
- Each domain has its own floor. Certification requires all six decision rules to be met. Rule 2 requires each of Structure, Test, Engage and Perform to be at or above 2.0 (Section 12.2). Failure of any one rule results in Not Certified.
- A strong domain does not replace a weak one. "Strength in one domain does not offset weakness in another" (Section 12.1). Domain scores are not averaged away into a single number for the pass decision. The composite is a separate check: rule 3 needs the mean of all nine dimension scores to be at or above 2.3 (Sections 12.2, 12.2.1).
- The candidate sees all four. Every candidate receives the four domain scores, every dimension score, the band and the confidence level. This is provided whether or not the candidate certifies (Section 12.5).
Where the name comes from
STEP takes its letters from the four domains: Structure, Test, Engage and Perform. Section 5 states that the model "is derived from the STEP judgment construct already canonical in AIR APAC's simulation methodology".
What the Standard says it does not yet know
Section 5.1 calls nine dimensions "a reporting hypothesis, not an established factor structure". The evidence programme checks whether the dimensions separate from each other in practice. The Standard states that the model "can shrink to four domains without breaking the certification decision rules". It adds: "It will shrink if the evidence says so."
What this does not claim
- It does not claim that the four domains are proven to be separate. Section 5.1 treats the structure as a hypothesis that the evidence programme tests.
- It does not claim that one domain matters more than another. The decision rules set the same floor for all four.
- It gives no level, band or score for any person. A result comes only from an Assessment under the Standard.
- It is not preparation material for an Assessment.
AIR APAC publishes the standard for judgment in AI-enabled work, certifies people against it through its official instrument, and stands behind every result.