Why are there four domains?
AIRS-J/1 groups its nine dimensions into four domains: Structure, Test, Engage and Perform. This page answers what each domain groups, why the Standard groups the dimensions this way, and what the grouping does in a result.
Judgment is how well a person decides once AI has given the answer. AIRS-J/1, the Standard for judgment in AI-enabled work, assesses it on nine dimensions. Section 5 of the Standard places those nine dimensions in four domains. The answers below quote the Standard and give its section numbers. Each domain gets one line here. The nine dimensions themselves are set out on the nine dimensions page.
What are the four domains?
Section 5 names four domains and gives each one a short line. The Standard's line is in quotation marks. The plain words after it are AIR APAC's explanation, not the Standard's wording.
- Structure: "Framing the problem before reaching for a tool." In plain words: the goal, meaning what the work has to achieve.
- Test: "Interrogating evidence, including machine-generated evidence." In plain words: the evidence, meaning what shows that a claim or an answer is supported, or wrong.
- Engage: "Authority, people, and being seen to act." In plain words: agreement, meaning who must agree and who must know.
- Perform: "Deciding proportionately, and owning the decision." In plain words: responsibility, meaning being answerable for the action and what follows.
Why does the Standard group nine dimensions into four domains?
Section 5 says where the grouping comes from. The model "is derived from the STEP judgment construct already canonical in AIR APAC's simulation methodology". Section 3 defines a domain as "One of four groupings of dimensions under the STEP structure".
Section 5.1 explains how the two levels relate. Nine dimensions is "a reporting hypothesis, not an established factor structure". The model "is designed so it can shrink to four domains without breaking the certification decision rules". In plain words, the decision rules would still work if the model had only the four domains. The Standard gives no other reason for the grouping, and this page does not add one.
Does each domain count on its own in a result?
Yes. A domain score is "the arithmetic mean of its dimension scores" (Section 12.2.1). To certify, each of the four domain scores must be at or above 2.0 (Section 12.2, rule 2). The rules are non-compensatory: "Strength in one domain does not offset weakness in another" (Section 12.1). The other decision rules are on the certification page.
Why does Test have three dimensions and the other domains two?
Section 5 lists three dimensions under Test: T1 Evidence Verification, T2 Assumption and Alternative Testing, and T3 Risk and Failure Anticipation. Structure, Engage and Perform have two dimensions each. The Standard gives no reason for this count, so this page does not offer one. Section 5.1 does name T1 and T2 as a known separation risk, because "both interrogate inputs".
The count has one effect on the composite. The composite is "the arithmetic mean of all nine dimension scores" (Section 12.2.1). Three of the nine scores in it come from Test. The domain floor treats every domain in the same way: each must reach 2.0 (Section 12.2, rule 2).
Could the number of dimensions change?
Yes, if the evidence shows it. Under the monitoring rule in Section 5.1, the evidence programme checks whether the dimensions measure different things. "A dimension that does not separate from another in its domain is merged in the next major version, and this Standard says so". The Standard names three known separation risks in advance: T1 and T2, S1 and S2, and E1 and E2. Section 5.1 ends: "It will shrink if the evidence says so."
What does a candidate see for each domain?
Every candidate receives "the four domain scores" (Section 12.5). The same report gives every separately reportable dimension score, the band, the confidence level and the evidence cited for each observation. "This is provided whether or not the candidate certifies."
Is a "domain" here the same as a field of work, such as finance or engineering?
No. In the Standard, a domain is one of the four groupings of dimensions (Section 3). Knowledge of a professional field is a different sense of the word. Section 2 lists "Domain knowledge in any professional field" as out of scope. The Standard assesses how a person decides, not what they know (Section 2).
Are the four domains the same as the parts of the AI Readiness Index?
No. The AIR APAC AI Readiness Index rates organisations on public signals (Section 16.1). The Standard calls the Index's parts "Index signal components", so that the names do not clash (Section 3, naming note). "The two models share no structure and are not convertible". A certification confers no Index effect (Section 16.1).
What happened to the earlier ten-domain model?
It is retired: the Standard supersedes all earlier competency structures (Section 5), and no historical score converts to an AIRS-J/1 result (Annex B.3).
Where does the name STEP come from?
STEP is made from the first letters of the four domains: Structure, Test, Engage and Perform.
What this does not claim
- It does not claim that the four domains carry equal weight in the composite. The composite is the mean of all nine dimension scores (Section 12.2.1).
- It does not claim that one domain matters more than another.
- It does not claim that nine dimensions or four domains are a proven structure. Section 5.1 calls nine dimensions "a reporting hypothesis".
- It gives no reason for the grouping beyond the reasons the Standard states.
- It gives no level or result for any person.
- 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.