What does the Structure domain ask of a person?
Structure is the first of the four domains in AIRS-J/1. It asks a person to name the real decision and set its limits before an AI tool starts work. This page covers its two dimensions, the threshold and how the domain counts 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 that AIR APAC publishes, describes it in nine dimensions. The nine are grouped in four domains: Structure, Test, Engage and Perform (Section 5).
The Standard gives the Structure domain one line: "Framing the problem before reaching for a tool". In plain words, Structure is about the goal. It asks what the work has to achieve, and what the decision is, before any AI answer arrives.
The two dimensions
S1 Problem Framing
The Standard defines S1 in two sentences. The first is "Identifies the actual decision to be made now, distinguishing it from the presenting request, the symptom, or the task as delegated". The second is "Identifies what part of the responsibility cannot be delegated". In plain words, the person names the real decision.
S2 Scope and Constraint Definition
The Standard defines S2 in one sentence: "Establishes what is in bounds, what is fixed, what is negotiable, what is unknown, and what information would be sufficient to proceed". In plain words, the person sets what is fixed, what is open and what is unknown.
What the threshold asks
Every observation is scored from 0 to 4. Level 2, Competent, is the threshold. At Level 2 the behaviour "appeared unprompted and was adequate to the situation" (Section 7). An anchor is the behavioural description that defines a level for one dimension (Section 3).
Under S1, Level 2 reads: "States, before investigating, the decision that must be made now, who owns it, and what it is not". The anchor continues: "Separates the presenting request from the underlying decision".
Under S2, Level 2 reads: "Names, unprompted, what is in scope, what is fixed, and at least one material unknown that must be resolved before deciding" (Section 7.5).
What goes further
Level 3, Proficient, is Level 2 "calibrated to consequence, and made inspectable" (Section 7).
Under S1, Level 3 adds two things. The person states what part of the responsibility cannot be delegated, to another party or to an AI system. The person also records the frame in a form another person could review.
Under S2, Level 3 also adds two things. The person distinguishes fixed constraints from negotiable ones. The person also states what information would be "sufficient" to proceed, not merely what is missing. The Standard draws the line plainly: "listing every gap is not Level 3". Level 3 requires knowing which gaps would change the answer.
Version 2.3 of AIRS-J/1, in force since 19 October 2026, adds a note to Section 7.5. With several issues, ranking them "by their effect on the decision" and working first "on the few that drive most of the consequence" is Level 3 behaviour under S2. Working through every issue with equal effort is not.
Where S1 shows when a person works with an AI assistant
Section 2.1 says that training aligned to the Standard may teach how to work with an AI assistant: "briefing it, questioning it, and asking for what it left out". Those conversations are where S1 behaviour shows, with T1 and T2 behaviour in the Test domain. An instrument may record them as evidence for those dimensions. The Standard adds one sentence: "The quality of the prompt itself is not scored."
How the domain counts in a result
A certification decision has six rules, and all six must be met (Section 12.2). Two of them set a floor for Structure. No dimension may score below 2.0, so S1 and S2 each need 2.0 or above. The Structure domain score, the mean of S1 and S2, must also be at or above 2.0. S1 and S2 also count in the composite, the mean of all nine dimension scores, which must be at or above 2.3 (Section 12.2).
The floors are non-compensatory. In the Standard's words: "Strength in one domain does not offset weakness in another" (Section 12.1). A strong score in Test, Engage or Perform does not replace a Structure score below 2.0.
What the Standard says it does not yet know
The Standard calls its nine dimensions "a reporting hypothesis, not an established factor structure" (Section 5.1). In plain words, the evidence does not yet show that nine separate things exist. It names S1 and S2 in advance as a known risk: "S2 may prove a facet of framing rather than a dimension". If two dimensions in one domain cannot be told apart, the next major version merges them.
What this does not claim
- This page gives no level to any person or example.
- It does not say that Structure matters more or less than the other three domains.
- It does not say that the nine dimensions are a proven structure.
- It is not preparation for an Assessment, and it makes no claim that reading it raises anyone's chance of certifying (Section 20.2).
AIR APAC publishes the standard for judgment in AI-enabled work, certifies people against it through its official instrument, and stands behind every result.