One question to ask about any AI investment, before you sign anything. It takes about twenty minutes and it costs nothing.
The question
Does this still make sense if the hype disappears completely, and the price of the technology moves ten times, in either direction?
A higher price tests the economics. A lower price tests what happens when competitors can buy the same capability more cheaply. Both are useful ways to examine the assumptions behind a proposal.
If the case depends on the current price, make that dependency explicit.
1Start with the business case
Describe the work, the proposed change, and the expected benefit in one paragraph. Include a number you can explain. Use the example or guided prompt if you need help getting started.
The prompt asks about the work and helps you identify the figures you need. If a number is missing, measure it or label it as an assumption.
Bring the paragraph back here to examine its assumptions.
2Strike the hype
Click any sentence whose persuasive force depends on the technology being new, exciting, urgent, or something competitors are doing. Set those sentences aside so you can examine the underlying benefit.
3Four questions underneath it
4Your assessment
5Ask for another assessment
An assistant can help find assumptions you have overlooked. Ask it to examine the case, then check its reasoning and arithmetic yourself. A colleague who knows the work can provide a useful independent review.
You can compare responses from both assistants. Treat disagreements as questions to investigate.
WHAT A GOOD ANSWER LOOKS LIKE
It identifies unsupported claims, shows the arithmetic, and separates evidence from assumptions. It should explain what would change its conclusion. Agreement or disagreement alone does not tell you whether the assessment is useful.
The part worth keeping
An assistant can produce persuasive arguments for and against the same project. That makes it useful for exploring a decision, provided you check the claims behind each argument.
A confident answer can still contain incorrect facts or arithmetic. Ask what evidence supports it, which costs it includes, and what information is missing.
Use the comparison to decide what to measure next. The decision should rest on the work, the numbers, and the people responsible for the result.
A LIMIT OF THE EXERCISE
It is easy to be generous with your own assumptions. I have talked myself into projects this way. Ask someone who knows the work and can question the proposal to review it with you.
No signup. No email. Nothing you type is sent to Joust, and none of it is stored. If you choose to use an assistant: the buttons that open your case in Claude or ChatGPT carry it inside the link, so it goes to whichever assistant you choose. Review what you include before opening those links.
The Survival Test. A free tool from Joust.joustagency.com/tools/survival-test/
