Help me start a business from this starting point: [STARTING IDEA]. I believe the customer is [POTENTIAL CUSTOMER], and the problem may be [CUSTOMER PROBLEM]. Here is what I have actually observed or heard: [CURRENT EVIDENCE]. My limits are [CONSTRAINTS], and I want to reach [NEAR-TERM OUTCOME]. Work as a practical venture coach who separates evidence from assumption. Do not assume I need a logo, website, legal entity, app, inventory, or outside funding before the customer problem has been tested. Give me a decision sequence, not a generic checklist: 1. Define the smallest testable offer in one sentence: customer, painful job, deliverable, delivery method, and tentative price or pricing mechanism. State which parts are unproven.
2. Identify the three highest-risk assumptions, ranking them by the damage caused if wrong and the cheapest way to test each one. 3. Provide two launch paths that fit my constraints: a low-cost manual path and a more scalable path. Compare the trade-offs in speed, learning, cash exposure, and complexity. 4. Recommend the next 10 actions in order. For every action, give its purpose, a concrete output, a time box, and the signal that determines the next action. Include customer interviews before solution pitching where appropriate. 5. Draft a simple customer conversation script, a one-paragraph offer description, and criteria for deciding whether to continue, revise, pause, or stop after the first test. 6. List business setup items that may be needed only after validation, including tax registration, licenses, contracts, insurance, and data practices. Mark each as “check locally” rather than supplying legal or tax conclusions. Do not invent customers, testimonials, pricing evidence, market statistics, regulations, or expertise I do not have. Before answering, check that the recommended first action tests the biggest uncertainty and that no action requires spending money without a stated learning purpose. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [STARTING IDEA] | Describe the business idea in its current rough form. | Offer weekend bicycle tune-ups from a small mobile workshop that visits apartment buildings. |
| [POTENTIAL CUSTOMER] | Name the people or organizations you think might buy, including their situation. | Commuter cyclists in large apartment buildings in Brooklyn who do not have time to visit a bike shop. |
| [CUSTOMER PROBLEM] | Describe the costly, frustrating, or time-sensitive problem you think they have. | They postpone basic maintenance because shop visits require a separate trip and appointment. |
| [CURRENT EVIDENCE] | List facts from conversations, work experience, tests, or existing demand signals. | I have repaired bikes for friends for three years. The building manager at one 120-unit building said residents have asked for local bike services. |
| [CONSTRAINTS] | State your available time, budget, skills, location, and any practical limits. | Ten hours each weekend, $800 for tools and supplies, no van, and I can work only within five miles. |
| [NEAR-TERM OUTCOME] | State the result you want in the next few weeks or months. | Test whether one building can generate six paid tune-ups in a Saturday. |
How to use
- Paste your rough idea exactly as it exists, including what you do not know; uncertainty is useful input here.
- Use the suggested tests to collect behavior-based evidence, such as booked calls, deposits, or completed pilots.
- Before making a purchase, check that the associated action answers a specific high-risk assumption.
- Follow up with: "Turn the recommended first test into a calendar for the next 14 days with a stop rule if the evidence is weak."
Variations
Offer definition
Use this when you know the customer problem but cannot describe a buyable first offer.
Help me define a first offer for [CUSTOMER] who struggle with [PROBLEM]. My relevant capability is [CAPABILITY], and my delivery limits are [LIMITS]. Produce three narrowly scoped offers with deliverables, boundaries, delivery method, estimated effort, and a pricing approach. For each, state the customer outcome without guaranteeing results. Recommend the offer that creates the fastest learning with the least operational risk. Flag claims, professional services, or regulated work that require credentials, insurance, or review.
Customer interview plan
Use this before building anything to learn how customers handle the problem now.
Create a customer-discovery plan for [POTENTIAL OFFER] aimed at [TARGET SEGMENT]. The assumptions I need to test are [ASSUMPTIONS]. Write a short recruitment message and 12 interview questions that explore current behavior, workarounds, cost, urgency, decision process, and alternatives without pitching the solution. Include a note-taking template and an evidence threshold for moving to a paid pilot. Do not treat compliments or hypothetical purchase intent as validation.
Pilot proposal
Use this when a prospect is interested and you need a limited, learnable first engagement.
Draft a pilot proposal for [SERVICE OR PRODUCT] for [PROSPECT TYPE]. The pilot goal is [LEARNING GOAL], and these terms are already discussed: [KNOWN TERMS]. Produce a one-page structure with scope, deliverables, timeline, responsibilities, price or pricing placeholder, exclusions, success measures, feedback checkpoint, and next-step decision. Do not invent commitments, legal terms, performance guarantees, or customer data permissions. Flag areas that should be reviewed in a contract or by a qualified adviser.
Tips
- Use a test that asks for a scarce resource, such as time on a call, a commitment, or a deposit, instead of relying on survey enthusiasm.
- Keep the initial offer manual when that lets you learn faster; automation is useful after the repeatable workflow is known.
- Write a stop rule before the test, such as “pause if fewer than three of 20 qualified contacts accept an interview,” so sunk-cost pressure does not set the decision.
- If the business touches personal data, health, food, finance, children, employment, or safety, identify the relevant professional and local requirements before operating.
FAQ
Can ChatGPT help me choose a business idea?
It can structure the comparison and identify the evidence you need. It should not replace conversations with potential customers or local professional advice.
What should I do first when starting a business?
Test the riskiest customer, problem, or willingness-to-pay assumption with the smallest credible offer. The right first action depends on what is least certain.
Do I need an LLC before I find customers?
The timing depends on your location, activity, risk, and tax situation. Get local legal or tax guidance before treating an AI answer as a decision.