# Build a business case without inventing an ROI number

Original 5HO practice material. Fictional example. Reviewed 5 October 2026.

## Useful result

You'll finish with a business case that states what is known and what must be tested.

A business case can be useful before you have a glossy savings number. Show what you measured, what you assumed and what would change the decision. Precision is not the same as evidence.

## Set up

Allow 25 minutes. Use a calculator and your trial log. If you have no measurements, use the invented figures here and label the result as a worked example.

## Practice input

Observed over 5 runs: total 30 minutes saved after review. Assumed frequency: 20 runs a month. Proposed subscription: AUD 30 monthly. Staff time valued for planning at AUD 40 an hour. No staff costs will actually be removed.

## Do the task

1. Calculate the measured average and a separately labelled monthly scenario.
2. Compare the cost with the value of capacity, not automatically with cash savings.
3. List adoption, quality and frequency assumptions before recommending a purchase.

## Prompt

Build a short decision note from these figures. Show observed average time saved, projected monthly capacity, subscription cost and assumptions. Distinguish released time from a reduction in spending.

## Refine

Calculate the task frequency needed to cover the subscription under the stated assumptions. Do not claim that covering cost proves the tool is worthwhile.

## Expected checks

Average saving is 6 minutes; 20 runs suggests 2 hours or AUD 80 of capacity value. At AUD 4 per run, 8 whole runs exceed the AUD 30 subscription. These are conditional estimates, not payroll savings.

## Save

Save the decision record with the trial log and a review date. Check the assumptions after the first paid month.

## Sources

- Writing clear prompts: https://help.openai.com/en/articles/10032626-prompt-engineering-best-practices-for-chatgpt