From Remediation to Prevention

Purpose of this document. A remediation project corrects what has already been paid. It does not change the conditions that produced the error. This page sets out the options available to close that gap, in the order we would normally recommend considering them.

Why the exposure remains after remediation

The errors identified in this review — including leave accrual and payment calculations, and rate application — were not caused by a single mistake. They were produced by an interpretation encoded into a system at a point in time, applied confidently every cycle since, with nothing independently checking the result.

Three conditions typically remain in place after a remediation project closes:

The interpretation is still the only one. Payroll calculates and validates using the same configuration. There is no second opinion on the answer.
Detection is still reactive. The next error will surface through a complaint, an audit, or the next review — after the money has moved.
The knowledge is still personal. Rules and edge cases live with individuals rather than in a system, and leave when they do.

Under the current wage compliance regime, the defensible position is no longer "we corrected it when we found it." It is evidence that pay was checked, every worker, every cycle — and a record of what was checked and why.

Options for consideration

Presented from lowest to highest investment. Most organisations begin at Option 2 or 3.

The options rise from lowest to highest investment. The pivot is Option 3 — where checking moves before the pay run, from correction to prevention.

Option 1 — Scheduled manual assurance

Formalise the spot-check methodology used in this review into a documented, repeatable internal procedure, run quarterly against a defined sample.

Addresses: recurrence of known error patterns.
Limits: sample-based, labour-intensive, dependent on the reviewer's expertise, and produces limited evidence.
Best where: the workforce is small, awards are simple, and appetite for system change is low.

Option 2 — Encoded rule checks over payroll output

Convert the award, agreement and policy interpretations confirmed during remediation into executable checks that run against payroll output each cycle, flagging variances against expected rates, penalties, span limits, break entitlements, overtime thresholds and leave accrual.

Addresses: silent recurrence and compounding at scale — every worker rather than a sample.
Limits: still detects after calculation, though before the next cycle compounds it.
Best where: payroll configuration is trusted broadly but has known weak points.

Option 3 — Pre-run (shadow) validation

Independently recalculate the pay period against separately encoded rules before export, and compare against what payroll is about to pay. Discrepancies are surfaced with the clause reference and the expected value attached; high-risk lines are held for review.

Addresses: the core structural problem — payroll currently marks its own work. An issue caught pre-run is a fix; caught post-run it is a correction; caught by the regulator it is a finding.
Best where: an error caught after payment is materially more expensive than one caught before it.

Option 4 — Upstream control of roster, time and approvals

Extend checking to the inputs: variance between rostered, worked and approved hours; missing or drifted clock events; unapproved overtime; and override authority captured by name and reason.

Addresses: the origin of most pay errors, which are input errors rather than calculation errors.
Best where: rostering, time capture and payroll sit in separate systems with manual handoffs.

Option 5 — Governance and evidence layer

Above the operational checks, maintain a register of pay-related policies, rules and risks, each mapped to a control, an owner, and a test result. Retain per-worker, per-period evidence of what was checked, against which rule version, with what outcome.

Addresses: the board and audit-committee question — "can we demonstrate we are paying people correctly?" — with a standing answer rather than an assertion.
Best where: there is director-level exposure, an audit committee, or a prior remediation on record.

Suggested sequencing

Horizon	Action
0–30 days	Document the interpretations confirmed during remediation, and the error patterns found, as a formal rule set. Assign owners.
30–90 days	Encode the rule set and run it retrospectively over 6–12 months of payroll to confirm scope and establish a baseline.
90–180 days	Move checking to pre-run. Extend to roster and time inputs where handoffs are manual.
Ongoing	Report exceptions and control test results to the audit committee each cycle. Re-version rules as awards and agreements change.

What to require of any solution considered

Regardless of the option or provider selected, we recommend the following be treated as non-negotiable:

Rules are versioned and citable. Every check must state which clause and which rule version produced the result.
Coverage is complete, not sampled. Every worker, every cycle.
Evidence is retained automatically. Inputs, rule, result, reviewer and timestamp — retrievable per worker, per period, without reconstruction.
Your interpretations remain yours. Where organisation-specific interpretations or agreements are encoded, ownership and portability should be explicit.
Validation is independent of payroll. A check that shares payroll's configuration inherits payroll's assumptions and will reproduce the same error.
Exceptions route with the fix attached. Flagging a variance without the expected value and clause creates work rather than removing it.

Prepared as part of the payroll review engagement. Options are presented for consideration and do not constitute a recommendation of any specific vendor.