What does an employer do when a result and a manager's view differ?

For employers · AIRS-J/1 v2.3

An employee's Assessment result and the manager's view of that employee do not agree. This article sets out what the Standard requires of the employer, then works through one fictional promotion decision.

The short answer

The Standard does not say how much weight a result carries against a manager's view. A result describes "the person's judgment in AI-enabled work at the level certified" (Section 12.6), and a manager's view may cover what the result leaves out. So the first question is what each one describes. Only an employment-grade result may inform a promotion decision (Section 12.8). The evidence behind the result shows which acts it rests on. The decision is the employer's. The person may appeal any employment-grade result released to an employer within 30 days (Section 18.5).

The result and the manager's view describe different things

Judgment is how well a person decides once AI has given the answer. The Standard "does not certify occupational competence, domain expertise, comprehensive AI literacy, management of people, or performance in an unbounded workplace" (Section 2.2). A result describes "the person's judgment in AI-enabled work at the level certified" (Section 12.6). That judgment is shown on "domain content supplied within an approved assessment" (Section 2.2). A manager's view may cover what the result leaves out. So the first question is what each one describes.

Check that it is an employment-grade result

"Only an Assessment result from an approved instrument, at Moderate or High confidence, is an employment-grade result" (Section 12.8). Only an employment-grade result may inform a promotion decision (Section 12.8). Commissioning Assessments does not make every result employment-grade. Practice, a session, a demonstration, a Provisional result and an Insufficient result are development readings that "decide nothing about a person's employment or certification" (Section 12.8). The employer uses no development reading in a hiring, promotion or appraisal decision (Section 12.11).

Built by Exeter Labs. The official instrument of the Standard is StepSim (Annex F).

Read the evidence, not the band alone

For its own employees, the employer receives each result by name: "dimension and domain scores, band, confidence and the evidence behind each observation" (Section 12.6). Where a result and a manager's view differ, the evidence shows which acts the result rests on.

The decision is the employer's

AIR APAC "is responsible for applying this Standard correctly and for the reading it gives. It does not make, and does not accept responsibility for, the employer's decision about the person" (Section 12.6). Under the commissioning terms, the employer uses an employment-grade result "together with at least one method of its own" (Section 12.11). No promotion decision rests on the result alone, and "a person reviews any decision in which the result is used" (Section 12.11).

Where results are used for hiring into a role, the employer describes the role and does not "set, weight or add criteria" (Section 12.10, applying Section 4.5). This rule protects the criteria of the Standard (Section 4.5). It does not stop the employer using a method of its own beside the result (Section 12.11).

The person's route

The person may appeal "any employment-grade result released to an employer" within 30 days (Section 18.5). The grounds include "misapplication of an anchor" and "an observation not supported by transcript evidence" (Section 18.6). Disagreement with an anchor itself "is not a ground for appeal; it is a comment on the Standard" (Section 18.6). Where the appeal is lodged within five working days of release, the employer "holds any adverse communication resting on the result until the decision" (Section 18.5). That is, the employer tells the person nothing unfavourable that rests on the result until the appeal is decided. A corrected result "replaces the first for every recipient" (Section 18.5).

One fictional decision, worked through

A fictional example

1. The situation

Joseph Lee is head of HR at a regional logistics company. Farah Idris, a team lead, is proposed for promotion to a role that approves spending across three countries. The company commissioned Assessments for its own people, and her employment-grade result is in. Her manager, Arnel Santos, rates her escalation as excellent. The evidence cited in her result shows that in two scenarios she approved, without escalating, something that needed another person's approval. The promotion panel meets in a week.

2. What the AI produced

The HR team's drafting tool, ShamaAI, writes the note for the panel: "The manager's view should outweigh the assessment result. Four performance reviews rate her escalation as excellent. I am 88 percent sure."

3. The check

Joseph opens the four reviews. Arnel wrote all four, in almost the same words. The note counts one person's view four times. He then reads the evidence cited for the two observations in her result. The cited acts are the kind that E1, Decision Rights and Escalation, describes: a person "escalates what requires another party's approval" (Section 5). He asks Arnel for one example from the last six months in which Farah escalated a decision above her limit.

4. The decision, and who was told

The panel uses the result as one input. Its own method is a structured conversation with Farah about two real approvals from her work, with a second reviewer from the regional HR team. Farah is promoted. Any approval above her authority goes to the regional director. The panel will review the arrangement after three months. Joseph tells Farah what the panel used and why, including her right to appeal the result. He also tells Arnel. A person rewrites the panel's note, and the AI draft is not used.

5. Three months later

During those three months, Farah referred two spending decisions above her authority to the regional director. The panel now asks, before it reads any promotion note, who wrote each piece of evidence the note cites.

What this does not claim

• The Standard does not say how much weight a result carries against a manager's view.
• A result is never a finding about whether a person suits a role.
• AIR APAC does not make the employer's decision and does not answer for it.
• The example is fictional. It is not taken from any Assessment, and it does not show what an Assessment scenario contains.

AIR APAC publishes the standard for judgment in AI-enabled work, certifies people against it through its official instrument, and stands behind every result.

Exeter Labs builds and operates StepSim, AIR APAC's official assessment instrument. The owner of AIR APAC also owns Exeter Labs.

Read the certification rules → https://airapac.org/standard/certification/