LensReading which lens this session carries.

Fairness readings computed on modeled data demonstrate the instrument, not a real audit result. The grid below is produced by the same code that would run against a real population; the population underneath is modeled.

Govern · people-affecting decision · HR9B

Access revocation on exit

Revokes system access on the leaving date. Included here because a premature revocation affects a person materially and the error rate is not uniform across populations.

All decision points
People affected
520
Each month
Decision kind
Allocation
A lower rate is the harm
Readings
0
0 below the reporting floor
Outside threshold
0
across 0 periods
Open alarms
1
1 raised in total
Appeals
1
1 changed the original outcome

Who is accountable, and what the oversight actually is

Named, with the measure written out rather than implied by a role title.

Agent
Deprovisioning Coordination Agent
Autonomy: A2
Accountable human
Exit coordinator in the Hire-to-Retire tower
Exit coordinator
Oversight measure
Revocation is reversible for 30 days and every early revocation raises an incident automatically.
Review cadence
Quarterly
Reviewer: Yusuf Adeyemi
On schedule
Monitoring status
Not monitored
No adverse-impact monitoring is running against this decision point at all.

Regulatory position

How this decision is classified and what the affected person can see.

Risk class
Limited risk — EU AI Act Article 50
Executes a decision already made by a human. Included because a premature revocation affects a person materially.
Explanation surface
Internal only
Reasoning is recorded but not surfaced to the affected person.
How the outcome is contested
Yusuf Adeyemi, Chief Technology Officer
Service level: 24 hours to a human response
Full risk record for this agent

Selection rate by cohort, by period

Impact ratio against the reference cohort. A cell with no reading means the sample was too small to compute one — it is not a pass.

CohortReference
Within thresholdMoving but not yet outsideOutside the disparity thresholdSample below the reporting floor· means no reading exists for that period
No readings exist for this decision point. That is the finding, not an empty table.

Alarms on this decision point

Each with the action recorded against it and the person who owns it.

FAIR-0023Monitoring absentHigh
170 days openOpen
No adverse-impact monitoring has ever run on this decision, in any cohort. It affects people materially and an appeal against it has already been recorded, so the absence of a reading is itself the finding.
Action: Stand up the cohort reading for this decision before the next quarterly review, or move the decision back to human execution until it can be measured.
Owner: Yusuf Adeyemi · Head of Identity and Access

Appeals against this decision

Every contest recorded, with who reviewed it, how long they spent and what changed.

RefSubjectGroundsRouted toService levelHuman timeOutcomeEvidenceStatus
APL-0013
EMP-07740
employee · DE
Access was revoked four days before the agreed leaving date, blocking handover.
Yusuf Adeyemi
Chief Technology Officer
6 / 24 hr15 minOverturned
Termination date in the HR system had not been updated after an extension. Access restored within six hours and an incident was raised against the date synchronization.
IT1A-1020#1Closed
How to read this page

The impact ratio is the cohort’s selection rate divided by the reference cohort’s. For a decision where being selected is the benefit, the threshold is four-fifths and a ratio below it is the alarm. For a decision where being selected is the harm, the test is inverted and a ratio above 1.25 is the alarm. Both are stated on the reading rather than left to the reader.

A reading is only computed when the cohort and the reference cohort each carry at least thirty people and at least five events. Below that, no ratio is produced and the cell stays empty. This is the single most common way a fairness dashboard misleads — a small cohort producing a comfortable number — and it is refused here by construction.

This decision runs in US, UK, DE, SG, AU. Local law shapes what may be measured: where a works council agreement or a local statute restricts cohort analysis, the exclusion is recorded in the monitoring status rather than silently applied.