Use it before you hire, buy equipment, change prices or drop a service. Works in ChatGPT or Claude.
The prompt
I'm choosing between these options: [list them]. Before you answer, ask me up to three questions you need answered to compare them properly. Then compare the options in a table by cost, risk, time to set up, and how easy each one is to undo. Then give me the strongest reason my favorite option could fail. My favorite is [option].
What it did when I ran it (24 Sep)
I ran it for a made-up roofing company that's turning down work: hire a second crew lead, subcontract the overflow, or stop taking jobs under $2,000. The favorite was hiring.
It asked three questions first: how much work is being turned down each month, what a crew lead costs all-in, and what margin you keep after paying a sub. The example answers: about $40,000 a month, about $75,000 a year all in with 6 to 8 weeks to be productive, and about 15% margin with one trusted sub.
Then it built the table:
Hire a second crew lead: about $6,250/month all-in. Moderate risk: fixed payroll and hiring risk. 6 to 8 weeks to set up. Easy to undo: medium-low.
Subcontract the overflow: keep about 15% margin, around $6,000/month. Moderate risk: quality and one trusted sub. Fastest to set up if the sub has capacity. Easy to undo: high.
Stop jobs under $2,000: little direct cost, but you lose the small-job revenue. Low operational risk. Immediate. Easy to undo: very high.
The strongest reason hiring could fail: you spend $75,000 a year, wait 6 to 8 weeks, and find the work you're turning away doesn't have enough margin or steady volume to keep that crew busy.
How to read it
Look at "easy to undo" first. When two options are close, start with the one you can reverse. Here, ChatGPT suggested using the trusted sub as a bridge: test whether the demand lasts before committing to a hire.