Skip to main content
Documentation
Documentation
Administrator Guide
Last Updated: 2024-09-20
Concept: Payroll Result Abnormalities

Concept: Payroll Result Abnormalities

This topic applies to:
  • Payroll for Australia
  • Payroll for Canada
  • Payroll for the UK
  • Payroll for the U.S.
What are Payroll Result Abnormalities?
Payroll result abnormalities are any results that, for 1 or more reasons, are atypical for a regularly run pay calculation. Workday uses machine learning to determine normal pay calculation and result patterns, based on historical pay results and current and historical worker data.
Pay Anomalies uses workers' payment information, based on their current and historical payroll results, to learn any historical abnormal payment patterns. It then uses those patterns to predict abnormal payroll results for the current pay period.
The machine learning model also accounts for historical data such as whether administrators marked similar pay results as normal or abnormal. Workday then identifies payroll results that violate those patterns and assigns the result a score, quantifying the likelihood that the result is abnormal.
When Workday identifies a payroll result as abnormal, it doesn't mean that the result is invalid. Rather, the result is atypical and might require additional attention. As the machine learning model gathers more data, Workday identifies abnormalities more accurately.
Abnormality Confidence
Confidence level is the model’s level of confidence in its prediction of a pay anomaly being either Normal or Abnormal. Possible values are: High, Medium, Low, or None. None might be applied when there are insufficient results to base a prediction on; for example, for a new employee.