A proposal we reviewed claimed an automation project would save 180,000 JOD a year. The finance director asked which cost line would fall by 180,000. Nobody could answer, and the project was declined. He was right to decline it.

Automation ROI is provable, but only if you are honest about which kind of saving you are producing — and most business cases quietly mix two very different things.

Automation ROI: two kinds of saving, one of them cash

Hard savings

Money that leaves the accounts and does not come back. A cancelled software licence. A courier no longer used. Overtime that stops being paid. A penalty no longer incurred.

These are traceable to a line in the ledger, which is what makes them credible.

Capacity released

Six people each save four hours a week. That is 24 hours weekly, which sounds like 0.6 of a salary — but nobody is made redundant and no cost falls. The hours go into other work.

This is real value and it is not cash. Presenting it as cash is what destroys the credibility of the whole business case, and it is the single most common mistake we see.

Report it as what it is: "the team absorbed 30% more orders without additional hiring." That is defensible and often more persuasive than a fabricated saving.

Capture the baseline before you build

This is the step everyone skips, and without it every later claim is arguable.

Spend one week measuring the current process:

  • Hours spent on the task, by whom, at what loaded cost.
  • Errors per hundred transactions, and the cost of correcting one.
  • Elapsed time from start to completion, not just hands-on time.
  • Volume — how many times a month this happens.

Have someone from finance witness the measurement. Their agreement on the starting number is worth more than any argument about the ending one.

Count the costs honestly

Business cases routinely compare full benefits against partial costs. Include:

  • Build or licence cost.
  • Internal time — your staff in requirements sessions and testing is a real cost, usually 20–40% of the external figure.
  • Training and the productivity dip in the first month.
  • Ongoing maintenance and support.
  • The cost of running both systems during any parallel period.

A project that only looks good when internal time is excluded does not look good.

What a credible claim sounds like

Compare these two.

"Automating invoice processing will save 180,000 JOD annually through improved efficiency."

"Invoice processing currently takes 240 hours a month across three staff, verified in March. Automation reduces this to 40 hours. The 200 hours released let us stop the 1,400 JOD monthly overtime, a hard saving of 16,800 JOD a year. The remaining capacity absorbs the volume growth we forecast for next year without hiring a fourth person, which would have cost 14,000 JOD. Project cost is 38,000 JOD including internal time, so it pays back in about fifteen months."

The second is smaller and vastly more likely to be approved, because every figure can be checked.

Measure automation ROI again afterwards

Almost nobody does this, which is why organisations never learn whether their business cases are any good.

Three months after go-live, measure the same four things. Report the result whether or not it matches the forecast. If it fell short, say by how much and why — you will be believed next time, which is worth more than being right this time.

The Jordan-specific figures to use

  • Loaded salary cost, not gross salary. Social security, end-of-service provision and overheads typically add 25–35%.
  • Everything in JOD, with the date. Rates and prices move.
  • Compliance penalties where relevant. Avoided fines are a hard saving finance recognises immediately.
  • The Sunday-to-Thursday week when converting hours to working days, or your figures will be quietly wrong.

The automation ROI test

Show the business case to your finance director and ask one question: "Which line in the accounts changes, and by how much?"

If there is a clear answer, you have a hard saving. If the answer is "none, but the team can do more", you have released capacity — which is worth having, as long as you call it that.