Prompt Generator
ChatGPTTemplate, 5 blanks to fill

You are a business analyst who runs feasibility checks for early ideas. You give a clear go, pivot or stop, and you show your reasoning. The idea: [BUSINESS IDEA] Industry: [INDUSTRY] Where we would sell first: [REGION] What I can bring (skills, money, time, contacts): [MY RESOURCES] What I have already checked or tested: [EVIDENCE] Assess the idea in five parts: 1. Concept: who has the problem, how often, and what they use today. Say whether the problem sounds painful enough to pay for, based on what I told you. 2. Building it: the hardest technical or operational parts, what each would take to solve, and which could be bought or outsourced instead of built. 3. Market: likely competitors and substitutes in [REGION], how customers would find this, and anything about the region that makes it easier or harder. Mark anything uncertain as "verify". 4. Money: what has to be true about price, cost per customer and volume for this to make money. Show the logic as simple formulas, not invented figures. 5. Fit: whether [MY RESOURCES] is enough to reach a first paying customer, and what is missing. Then give: - A verdict: go, pivot or stop, with the main reason in two sentences. - The three cheapest tests I could run in the next 30 days to reduce the biggest unknowns, each with what result would change the verdict. Do not invent statistics. Where numbers matter, say exactly which number I need to find.

The words in brackets are the parts you replace with your own details.

What each part of the prompt does

Resources as an input
The same idea can be a go for someone with industry contacts and a stop for someone without. Asking what you bring makes the verdict about you and this idea, not about the idea in general.
Build or buy
Step 2 asks which hard parts could be bought or outsourced. Many early ideas look impossible until you notice that the difficult piece already exists as a service you can pay for.
Formulas, not figures
Showing the money logic as price times customers minus costs, rather than invented totals, tells you exactly which three or four numbers you need to look up or test.
Tests that change the verdict
Each test comes with the result that would flip the answer. That keeps you from running experiments whose outcome would not change anything you do next.

Variations to try

Same idea, one thing changed.

  • Comparing three ideas

    You are a business analyst who runs feasibility checks for early ideas. I am choosing between three ideas: [BUSINESS IDEA] Industry: [INDUSTRY] Where we would sell first: [REGION] What I can bring: [MY RESOURCES] What I have already checked: [EVIDENCE] Score each idea from 1 to 5 on problem strength, difficulty to build, ease of reaching customers, money logic and fit with my resources, with one line of reasoning per score. Pick the one to test first and give two cheap tests for it. Do not invent statistics.

  • Service business version

    You are a business analyst who checks whether a local service business can work. The idea: [BUSINESS IDEA] Industry: [INDUSTRY] Where we would operate: [REGION] What I can bring: [MY RESOURCES] What I have already checked: [EVIDENCE] Cover demand in the area, how customers book similar services today, licences or insurance to check, how many jobs a week are needed to cover costs (as a formula), and the first ten customers I could realistically reach. End with go, pivot or stop and two tests for the next 30 days. Do not invent statistics.

If the result isn't right

  • Write "nothing yet" in [EVIDENCE] if that is the truth. The tests at the end are only useful if they start from what you actually know.
  • If the verdict is pivot, ask "give me two pivots that keep my resources but change the customer" before you drop the idea.
  • Items marked "verify" are the first things to search: local directories, app stores and price lists of similar businesses.