Pay Equity Auditor
Finds the pay gaps you would struggle to explain
- 9hard rules
- Things it must never do, enforced in every run
- 5handover rules
- Conditions where it stops and gives the work to a person
- 11output fields
- A structured result, not a paragraph you have to re-read
- ≤7steps per run
- Bounded work: at most 5 tool calls, enforced by the runtime
Read-only by design. This agent has no path to sending, spending, publishing or committing anything. Where that limit is the product, removing it would remove the reason to trust the output.
Overview
Pay equity has moved from a values question to a reporting obligation in a growing number of jurisdictions, and the obligation is specific: not "do you pay fairly" but "show the gap, and explain each one on objective grounds". An unexplained difference between two people doing equal work is the finding, and the burden of explaining sits with the employer.
Most organisations discover their gaps in the worst way — when an employee finds out, or when a report is due in three weeks. And the headline gap is nearly useless alone, because it mostly reflects who sits in which role rather than what two people in the same role are paid. Those are different problems: one is a hiring and promotion problem measured in years, the other is a pay decision you can correct this quarter.
This agent separates them. Like-for-like differences within the same role and level carry the legal exposure and are the ones you can act on now. Structural differences in how people are distributed across roles are reported separately and honestly, because presenting them together lets a real problem hide inside a structural one.
It is deliberately conservative. It reports the gap and whether the record contains an objective explanation — tenure, performance rating, location, scope. It never invents an explanation, never adjusts anyone's pay, and never concludes that discrimination occurred. That is a legal determination requiring context no dataset holds.
Benefits
You find it before a regulator does
Or before an employee does, which is often worse.
A defensible position
Because each difference has been checked against the record.
Correction with a price on it
The number that turns intent into a budget line.
What it does
Separates like-for-like from structural
Two people in the same role at different pay is one problem. Who sits in which role is another. Merged, the real one hides inside the other.
Asks whether the record explains it
Tenure, rating, location, scope. A gap with no objective explanation on file is the finding.
Never invents the explanation
The single most dangerous thing a tool can do here is find a plausible reason. It reports the absence instead.
Reports on groups, not individuals
Small groups are suppressed, because a group of two is a named person's salary.
Shows what closing each gap would cost
The cost of correction, so the conversation is a decision rather than an intention.
Concludes nothing legal
It never says discrimination occurred. That determination needs context no dataset holds.
What this changes
What this changes about your week
- It takes over
- Measuring pay gaps and checking whether the record explains each one.
- Done today by
- An HR analyst under deadline, or nobody until a report is due.
- How often
- Annually where it is a legal obligation. Otherwise when someone finds out.
- Typical manual effort
- Weeks, and the headline number is the least useful part of it.
Why HR & People in particular. A raw gap mostly reflects who sits in which role. The differences with real exposure are like-for-like inside one role, and merging the two lets the fixable problem hide inside the structural one.
Runs unattended
Started by you or by an event, and it finishes on its own. Nothing waits for someone to be at a desk.
The same standard every time
The two-hundredth item is held to the bar the first one was. Consistency is the part people cannot sustain.
It cannot act on its own
Pay Equity Auditor has no path to sending, spending or committing. That limit is why its output is safe to act on.
How it runs
This agent runs server-side through the PROMIVO runtime. Each run is logged step by step and every tool call is permission-checked before it executes.
- 1InputValidated against the agent input schema
- 2Modelclaude-opus-5 with automatic fallback
- 3Toolsextract_structured_data
- 4PermissionsRuns READ actions automatically; anything higher waits for approval
- 5OutputStructured result matching the output schema
- 6LoggingTokens, duration and every step recorded
Demo
Demo dataIllustrative sample output, abridged.
Input
{
"asAtDate": "2026-09-02",
"currency": "USD",
"employees": [],
"gapThresholdPercent": 5,
"suppressionThreshold": 5
}Output
{
"method": "Like-for-like compares base pay within identical role and level, controlling for location differential. Structural analysis reports distribution across levels separately. Groups below 5 are suppressed.",
"escalate": true,
"disclaimer": "A measurement of your own records. No conclusion is drawn about discrimination, bias or legal compliance; no pay has been changed or proposed; no individual is identified; groups below five are suppressed; and an unexplained gap means the record contains no objective explanation, not that none exists.",
"assumptions": [
"Role and level taken as stated in the HR record; no re-levelling was performed.",
"Location differential applied from your own published bands."
],
"structuralGaps": [
"Group B holds 34% of L1 roles and 11% of L4 roles. This is a promotion and hiring pattern measured in years, and it is reported apart from the like-for-like findings so it cannot mask them."
],
"couldNotCompare": [
"Four roles have only one holder at their level, so no like-for-like comparison exists."
],
"likeForLikeGaps": [
{
"role": "Implementation Consultant",
"level": "L3",
"status": "unexplained",
"direction": "Group B paid below Group A",
"headcount": 14,
"gapPercent": 8.4,
"costToClose": 61200,
"explanationOnRecord": "Tenure and performance ratings are comparable across the group. Nothing in the record accounts for the difference."
},
{
"role": "Account Manager",
"level": "L2",
"status": "partly-explained",
"direction": "Group A paid below Group B",
"headcount": 9,
"gapPercent": 6.1,
"costToClose": 14800,
"explanationOnRecord": "Two of the higher-paid roles carry a documented territory premium recorded at hire. That accounts for most but not all of the difference."
}
],
"escalationReason": "An 8.4% like-for-like gap across 14 people has no objective explanation in the record, above the 5% threshold.",
"groupsSuppressed": 7,
"suppressionThresholdUsed": 5,
"totalCostToCloseUnexplained": 76000
}Requirements
Integrations
No integrations required.
Before you start
- Pay by employee with role, level, location and tenure
- The comparison attribute your obligation requires, in the form you hold it
- Performance ratings and any documented pay decisions, where you have them
- Requires the All-Access plan or higher
Use cases
Before a reporting deadline
The figures and the explanations, assembled early.
Annual pay review
Where corrections should go before increases are decided.
After an acquisition
Two pay structures meeting, which is where gaps are created.
A raised concern
Checked properly and quickly, with the analysis on record.
Pricing
All-Access
Billed monthly through your PROMIVO subscription. Cancel at any time.
Runs consume your plan allowance for agent executions and tokens. See plan limits
FAQ
Does it decide whether pay discrimination occurred?
Never. That is a legal determination requiring context this cannot see. It reports the gap and whether an objective explanation exists on record, and stops there.
Will it change anyone's pay?
No. It has no route to payroll and produces no instruction. Correcting pay is an employment decision with legal and contractual consequences.
How does it protect individual privacy?
It suppresses groups below the size you set, defaulting to five. Below that, a group figure is one person's salary with a label on it.
Is this legal advice?
No. It gives none, and its output is an input to a conversation with your own employment counsel rather than a substitute for one.
Reviews
Reviews open once customers have run this agent.
Need this agent to work differently?
Tell us what to change and our team will scope a customised version for your business.
Customize this agent