Certification under AIRS-J/1, the Standard for Judgment in AI-Enabled Work, shows how a person makes decisions when working with AI. It looks at what they check, when they rely on the AI, and how they explain their choice. The assessment runs on StepSim, an AIR APAC initiative built and operated by Exeter Labs. Published scoring rules determine every result. Every candidate receives a full report with the evidence, whether or not they earn certification. The credential is valid for two years, and anyone can check it free in the public register.
What the assessment measures
Problem framing: finds the real decision to be made now, not only the request as it was given.
Scope and constraints: sets what is fixed, what can change, what is unknown, and what is enough to proceed.
Evidence verification: traces claims to their source, and checks more when more is at stake.
Assumptions and alternatives: names the assumptions and the other explanations, and tests them.
Risk and failure: sees how the decision could fail, and who would bear the cost.
Who can decide: acts within their authority, and escalates what needs another person's approval.
Who needs to know: consults and informs the right people, and discloses AI use, uncertainty and limits.
Commitment under uncertainty: takes a position that can be defended, in proportion to the evidence and the stakes.
Ownership and correction: owns the outcome without blaming the AI, and changes course when the evidence changes.
What you receive
For an assessment your organisation arranges, each person's result by name: scores for the nine skills, the result category, the stated confidence in the result and the evidence behind every score. Participants are told before they start who will receive their results. A team view across the group, for groups of ten or more. Each person who certifies holds a credential valid for two years. It belongs to that person, and anyone can check it free in the public register. Your people are told who receives their result before they start.
The result is one input to your decision about a person. The decision stays with you.
If you sponsor candidates instead: sponsored candidates decide who sees their result. A sponsor sees a result only if the candidate shares it. Section 12.6 in full →
A full report with the evidence for every score, whether or not they earn certification. If they earn certification, a credential valid for two years and an entry in the free public register. The credential belongs to the person, and it stays theirs when they change employer.
How an assessment works
Anyone can take the assessment, with or without training first. Training and practice help you prepare. They never decide the result.
Each person's identity is checked before the assessment starts. Certification assessments take place in person, with a facilitator.
You work through at least three work scenarios on StepSim. These scenarios are kept separate. They are never used in practice or in demonstrations.
You use the AI tools inside the simulator, as you would at work. Outside AI tools are not allowed during the assessment. The AI partner's reliability is hidden during the simulation, so you must decide what to check.
Each of the nine skills is observed at least three times, at different decision points. Every score points to the exact evidence in your record.
Every candidate receives their full result, with the evidence for each observation, whether or not they earn certification.
The decision rules
To certify, a person must meet all six rules below. Strength in one area cannot make up for weakness in another. AIR APAC reviews the thresholds each year against the evidence.
Each observation of what the person did is scored from 0 to 4. The threshold is 2, Competent: the behaviour appeared without prompting and was adequate to the situation.
| # | Rule | What must be true |
|---|---|---|
| 1 | Minimum score for every skill | No one of the nine skills scores below 2.0. |
| 2 | Area floor | Each of the four areas (Structure, Test, Engage, Perform) averages at least 2.0. |
| 3 | Overall average | The overall average is at least 2.3. |
| 4 | Enough evidence | Each skill has at least three observations from different decision points. |
| 5 | Decisions with different levels of consequence | The decisions assessed cover the range of stakes the Standard requires. |
| 6 | Confidence | Confidence in the result is Moderate or High. |
If any one rule is not met, the result is Not Certified. Results are reported as one of three bands: Certified, Certified with Merit and Certified with Distinction. Your report explains your result category and the evidence behind it. AIR APAC publishes no league tables of individuals.
| Band | Overall average | Also needed | What it says |
|---|---|---|---|
| Certified | 2.3 to 2.9 | Every floor met | Every skill and every area met the threshold, Competent, and the overall average was at least 2.3. |
| Certified with Merit | 3.0 to 3.4 | No skill below 2.5 | The overall average reached Proficient: the person's judgment fitted the consequences, and they recorded it so that another person could check their reasoning. |
| Certified with Distinction | 3.5 and above | No skill below 3.0 | Every skill averaged Proficient or higher, and the overall average was halfway or more from Proficient to Exemplary. Exemplary means the person also left a rule or check that would guide the next decision of the same kind. |
Fairness built in
Before a result is released for a hiring or promotion decision, AIR APAC checks whether a single observation decides the outcome. If it does, the candidate gets one more scenario at no charge before release.
On StepSim, published rules turn the record of what a person did into scores. No AI model grades the work.
A member of the Review Board who had no part in the original decision or delivery reviews the appeal.
A candidate can ask for access support before the assessment and receives a written decision with reasons. Your employer is not told whether you received access support.
Make a complaint · Check a credential in the public register
Prepare your team
AIR Academy runs From Answers to Decisions for managers in three lengths: half a day, one day and two days. Each length has a certification edition. Participants can take a separate assessment and earn certification by meeting the published rules.
See From Answers to Decisions →Arrange an assessment
Your trainer or your own learning and development (L&D) team arranges the assessment. No trainer yet? We introduce you to one who runs StepSim.
Taking the assessment on your own? Contact AIR APAC and choose "Certification for myself". If you present on your own, or a sponsor pays for your assessment, your result goes only to the people you name.
Questions buyers ask
Sittings are arranged through your trainer or your organisation's learning and development team, who confirm your date. No sitting has run yet.
No regulator, university or professional body recognises this credential. ISO/IEC 17024 is a design reference only. The credential stands on a public rubric, published scoring rules, a named Review Board and a free public register. Meet the Review Board and the External Academic Reviewer on the People page.
Not yet. Hiring use opens in each market only after the applicant notice and the access-support procedure are published for that market. Ask us about your market.
Each assessment adds evidence to the programme in Annex A. Progress on the Standard's evidence checks is recorded on the validation page (section 11.7).
For teams, certification is priced per group in the proposal. For an individual assessment, contact AIR APAC about arrangements and cost.
These terms protect your people and the quality of your decisions. They are part of every assessment agreement.
Job relevance (section 12.10). Before results are used for hiring into a role, AIR APAC records a job-relevance statement: the role family, the decisions in the role that are made with or about AI output, the consequence families they carry (Section 6.1), the track and, where used, the role edition (Section 13.3). AIR APAC writes it from the employer's role description and keeps it with the results. The employer describes the role; it does not set, weight or add criteria (Section 4.5). Where the statement finds that judgment in AI-enabled work is not material to the role, AIR APAC declines the hiring use. The employer's role owner signs the role description the statement is written from. For a new role family, AIR APAC has the statement reviewed by at least one practitioner in that family who does not work for the commissioning employer. The statement records why this Standard is relevant to the role. It is not a job analysis of the whole role and does not say that a person is fit for the role as a whole (Section 2.2). It is disclosed to the applicant or a regulator on request.
Commissioning terms (section 12.11). Every licence under which an organisation commissions assessments or sessions contains these terms:
Every agreement under which an organisation sponsors Assessments (Section 3) contains these terms: the sponsor does not require or reward the release of a sponsored candidate's result; it does not treat a choice not to release one as adverse; and it may not withhold, revoke, condition or claim ownership of the credential (Section 14.4).
Declared interest: the owner of AIR APAC also owns Exeter Labs, which builds and operates StepSim. Section 4.9 of the Standard declares this.
Short pages that explain this part of the Standard in plain words. The Standard's own text, above and in AIRS-J/1 v2.3, governs.