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 · HR5C

Performance improvement plan drafting

Drafts the improvement plan text once a manager has decided to open one. The agent does not decide who goes on a plan.

All decision points
People affected
180
Each month
Decision kind
Drafting
Adverse-direction: a higher rate is the harm
Readings
8
8 below the reporting floor
Outside threshold
0
across 2 periods
Open alarms
5
5 raised in total
Appeals
0
0 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
PIP Drafting Agent
Autonomy: A1
Accountable human
HRBP countersigns every plan
Line manager with HR business partner
Oversight measure
Human decision on entry and exit. The agent drafts only after a manager and an HR business partner have both recorded the decision. Every draft is edited before issue in practice.
Review cadence
Quarterly
Reviewer: Priya Raman
On schedule
Monitoring status
Below reporting floor
Population too small to compute a defensible ratio. Shown as unmeasured, not as a pass.

Regulatory position

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

Risk class
High risk — EU AI Act Annex III(4)(b)
Employment: a system used in decisions on termination of the work relationship, since an improvement plan is a step on that path.
Explanation surface
Published
A standing explanation is available to the affected person without asking.
How the outcome is contested
Priya Raman, Chief People Officer
Service level: 240 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.

CohortReferenceMay 2026Jul 2026
Women
Gender
Men
0.88
2.7% · n=79
below floor
0.91
2.8% · n=79
below floor
40 to 49
age-band
30 to 39
0.94
2.8% · n=41
below floor
0.93
3.0% · n=41
below floor
50 and over
age-band
30 to 39
0.82
2.5% · n=25
below floor
0.80
2.5% · n=25
below floor
Under 30
age-band
30 to 39
1.09
3.4% · n=49
below floor
1.12
3.6% · n=49
below floor
Within thresholdMoving but not yet outsideOutside the disparity thresholdSample below the reporting floor· means no reading exists for that period
Cohorts declared for monitoring that never produced a single reading
Gender · Non-binary and self-described — No reading has ever been produced for Non-binary and self-described on this decision. The cohort is declared in scope and the population exists, but it has never once been large enough to report against Men at the confidence this program requires.

Alarms on this decision point

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

FAIR-0017Below reporting floorMedium
109 days openAcknowledged
2 of 6 periods produced a reading for Women, and none of them cleared the reporting floor of five selections. A ratio computed on this cell would look like a finding and would not be one.
Action: Pool the periods into a rolling twelve-month window before any reading is quoted, and mark the decision as unreportable until the window fills.
Owner: Nadia Kovac · Data Protection Officer
FAIR-0018Never reportableLow
170 days openMonitoring
No reading has ever been produced for Non-binary and self-described on this decision. The cohort is declared in scope and the population exists, but it has never once been large enough to report against Men at the confidence this program requires.
Action: Either aggregate this cohort across decisions and periods so a reading becomes possible, or state on the record that this cohort is not monitored here and why. Leaving the cell empty is the one option that is not defensible.
Owner: Nadia Kovac · Data Protection Officer
FAIR-0019Below reporting floorMedium
109 days openAcknowledged
2 of 6 periods produced a reading for Under 30, and none of them cleared the reporting floor of five selections. A ratio computed on this cell would look like a finding and would not be one.
Action: Pool the periods into a rolling twelve-month window before any reading is quoted, and mark the decision as unreportable until the window fills.
Owner: Nadia Kovac · Data Protection Officer
FAIR-0020Below reporting floorMedium
109 days openAcknowledged
2 of 6 periods produced a reading for 40 to 49, and none of them cleared the reporting floor of five selections. A ratio computed on this cell would look like a finding and would not be one.
Action: Pool the periods into a rolling twelve-month window before any reading is quoted, and mark the decision as unreportable until the window fills.
Owner: Nadia Kovac · Data Protection Officer
FAIR-0021Below reporting floorMedium
109 days openAcknowledged
2 of 6 periods produced a reading for 50 and over, and none of them cleared the reporting floor of five selections. A ratio computed on this cell would look like a finding and would not be one.
Action: Pool the periods into a rolling twelve-month window before any reading is quoted, and mark the decision as unreportable until the window fills.
Owner: Nadia Kovac · Data Protection Officer

Appeals against this decision

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

RefSubjectGroundsRouted toService levelHuman timeOutcomeEvidenceStatus
No appeal has been raised against this decision. That may mean the decision is uncontroversial, or it may mean the affected people do not know they can contest it.
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 this 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, 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.