What does a manual workflow really cost?
A transparent worked example helps decide whether automation is worth trying. Change the assumptions before making a business case; the numbers below are illustrative.
First measure a real task, not a feeling
Pick one repeatable activity, such as copying invoice details from an email into a spreadsheet and then an accounting system. Observe several representative cases rather than estimating from memory. Separate routine time from exceptions, rework and approvals. Record how often the task occurs, who performs it and whether work peaks at particular times. If the task is rare, a better template may be more sensible than software.
Worked example: 25 items per day
Suppose a team handles 25 items on each of 22 working days. The current manual copying step takes three minutes per item. That is 25 × 3 = 75 minutes per day, or 27.5 hours per month. At an illustrative fully loaded labour cost of €28 per hour, the activity consumes approximately €770 of capacity each month. This is capacity cost, not necessarily cash that will be saved.
Estimate the realistic improvement
A proposed workflow removes 60% of the copying effort, leaving exceptions and human review. The gross capacity saving is 27.5 × 0.60 = 16.5 hours per month, worth about €462 at the assumed rate. But the replacement process has ongoing costs: suppose a €49 monthly licence and three hours of maintenance and monitoring at €28 per hour (€84). The estimated net monthly capacity benefit becomes €462 − €49 − €84 = €329.
Calculate payback with honest caveats
If the one-time implementation cost is €1,800, the simple capacity-benefit payback is €1,800 ÷ €329 ≈ 5.5 months, provided the assumed savings recur. This excludes financing, downtime, tax, training and failure risk. It also assumes freed time can be redeployed productively. If nobody can use that time, there may be no corresponding cash saving. A sensitivity check at 30% improvement instead of 60% gives only €231 gross benefit, and €98 net after the same ongoing costs.
Count quality and risk separately
Automation may improve audit trails, reduce transcription mistakes and clarify handoffs; it may also create new risks, such as incorrect field mapping or silent failures. Do not inflate a benefit estimate by counting time savings twice as both productivity and headcount savings. For quality improvements, record the current error rate, cost per incident and evidence that the change actually improves performance.
Use a small pilot before scaling
Test the proposed workflow on a bounded dataset that includes normal cases, missing information, duplicates and unexpected formats. Make sure a person can intervene when the automation cannot decide safely. Record who owns access rights, logging, exception handling and recovery after failure. Compare the pilot results against the original baseline using the same definitions.
A six-question decision checklist
Can we document the present process in plain language? Do we know monthly volume and time per item? Which cases require judgment? Who maintains the solution? Are data protection and permissions adequate? How will we prove the benefit after 30 and 90 days? If these answers are missing, start with workflow discovery, not procurement.
Next step
Use the result as a starting hypothesis, not a guaranteed return. If your company has a repeated administrative workflow, describe its steps, frequency and main exception at /request/?topic=automation. The aim is to reduce friction before adding software.