One-page project business case with Claude

A business case asks management for a decision, and the decision gets made on one page. Claude writes that page well from your calculation: the proposed decision, a table of metrics, the comparison with doing nothing, arguments and risks. The calculation is where the work is, and the whole task takes about an hour. What Claude tends to add, and you should remove, are benefits nobody can measure.

Calculate first

Describe the project in plain terms: what you buy or implement, what it costs up front and each year, which costs fall or which revenues grow. Then calculate in a spreadsheet the yearly effect, the payback period and net present value. The discount rate for NPV comes from the finance director; using your own invites an argument about the rate instead of the project.

Add the "do nothing" alternative with its own cost. This is the part most business cases skip, and it is often the strongest argument. Doing nothing about an ageing system is rarely free: rising maintenance, manual workarounds, the risk of failure during peak season.

The prompt

Role: you are a company economist preparing a business case for a management decision. Context: project: {{description}}. Cost: {{investment and annual costs}}. Effect: {{savings or revenue by year}}. Discount rate: {{rate}}. Task: write a one-page business case with the cost, effect, payback period, NPV and a comparison with the "do nothing" alternative. Use only the numbers provided. Format: the proposed decision in one sentence, then a table of metrics, three arguments, two risks and what we do about each. No long dashes.

The first line of the output, the decision in one sentence, is what the meeting will discuss. If it reads "consider implementing", rewrite it as a specific decision: "approve purchase of the warehouse scanning system with rollout in the second quarter".

Strip the soft benefits

Claude adds benefits such as "improved employee morale", "stronger customer loyalty" and "enhanced brand image". They sound reasonable and cannot be checked, and a management team that sees them starts discounting the numbers too. Delete every benefit that is not in your calculation. If you believe a soft benefit is real, measure a proxy for it or leave it out.

Then check every number in the text against the spreadsheet. Payback period and NPV are easy to transpose between scenarios, and a case where one number on the page does not match the attached calculation loses the room.

From page to slides

For the decision meeting, turn the page into three slides: the decision and the metrics table, the comparison with doing nothing, the risks and mitigations. Keep the numbers identical to the page.

What to do

1. Describe the project: what you buy, what it costs, which costs fall or revenues grow.
2. Calculate the yearly effect, payback period and NPV, with the discount rate agreed with the finance director.
3. Add the "do nothing" alternative and its cost.
4. Give Claude the calculation with the prompt.
5. Delete unmeasurable benefits and check every number against the table.
6. Turn the text into three slides for the decision meeting.

Ready for the meeting

Management sees the cost, effect, payback period and alternative on one page, and every number can be reproduced from the table.

Plans

Business cases are short and fit the free plan. Pro at 20 USD per month (17 with annual billing) adds Projects and extended reasoning, useful when you revise a case several times; Team at 25 USD per seat with annual billing keeps team data out of training by default. See Claude pricing, the economist's version of this task and Claude vs Copilot in Excel for the calculation side.