A payroll cycle involves collecting all the data required to process payroll and issue paychecks for a given time period (such as a biweek or month). Each payroll cycle requires participation by many actors, including individual employees, their supervisors, campus/division financial offices, a central payroll office, and central IT. Employees and departments report information such as hours worked (for hourly employees), reimbursements requested, lump sum payments, and leave taken (leave accrual is calculated as part of the payroll cycle). Other actors provide approvals along the way.
Suppose that currently all this reporting is primarily paper-based, up to the point at which central payroll keys in hours, reimbursements, leave, etc., into an ERP. For example, in the case of hourly employees, departmental timesheets are centrally printed, distributed to departments (and from there to sub-departments), completed, approved, locally archived (copies made), transmitted back to the central payroll office, and finally keyed in. Once payroll data is complete, there is additional participation by central IT to "run" payroll and print checks (for employees not receiving direct deposit).
Automation might make this process less expensive by reducing costs such as: printing, delivery, repeated data entry, time delays, and/or errors. These gains would need to be compared to the cost of automation.
Not every aspect of this scenario is a workflow problem. For example, reducing the amount of paper (and paper-handling) involved would start with an investment in self service functionality that delegates data entry to individual employees and integrates with the ERP that processes payroll. Also, delegation to individual employees could not succeed without detailed knowledge of employees and supervisory relationships such as would be found in an ERP containing human resources data.
Nevertheless there are distinct workflow issues: