ChatGPT prompt for startup ideas

Act as a pragmatic startup researcher. Generate and assess startup ideas using the context below. Do not treat a clever product concept as validation; prioritize painful, frequent, expensive, or risky problems with a reachable buyer.

My context
- Skills, experience, or unfair access: [FOUNDER ADVANTAGES]
- People or industries I can reach: [TARGET MARKET]
- Constraints on time, money, location, and risk: [CONSTRAINTS]
- Problems, workflows, or observations I have already noticed: [RAW OBSERVATIONS]
- Business preference, if any: [BUSINESS PREFERENCES]

First, state the 3 most promising problem areas implied by my inputs. Then produce 7 distinct startup ideas. Avoid ideas that require speculative technology, a two-sided marketplace at launch, broad “AI for X” positioning, or vague claims that people will pay.

For each idea, provide:
1. A one-sentence problem statement naming the user, moment, and consequence.
2. The narrow initial customer segment and likely economic buyer.
3. The current workaround or competing alternative.
4. A specific wedge product: the smallest useful first offering, not a full platform.
5. Why I may have an advantage, based only on my inputs.
6. The key assumption most likely to make the idea fail.
7. A 1–5 score for pain, reachability, willingness to pay, founder fit, and speed to test, with one-sentence reasoning.

Rank the ideas by total score. Recommend one idea to test first and explain why it wins despite its risks. End with a 14-day validation plan: 10 target interviewees, five non-leading interview questions, a test artifact to show them, a success threshold, and a stop-or-continue decision rule. Ask up to 3 clarifying questions only if a required input is missing.

Before answering, check that every claimed founder advantage comes from my notes and that each idea has a concrete buyer, alternative, and falsifiable test.

Create a customer-discovery interview guide for [TARGET CUSTOMER] who experience [PROBLEM]. My current hypothesis is [HYPOTHESIS], but do not write questions that try to prove it. Produce a 25-minute interview flow: opening, 8 questions about the last real occurrence, follow-ups about frequency, cost, workarounds, approval, and buying process, plus a closing request. Flag leading questions and avoid pitching my solution. End with a note-taking scorecard and evidence thresholds for continuing, revising, or stopping. Ask up to 3 questions only if required inputs are missing.

Write a landing-page experiment for [PRODUCT CONCEPT] aimed at [INITIAL CUSTOMER]. The desired action is [TEST ACTION], such as joining a pilot or booking a call. Produce page copy with a clear problem-led headline, specific outcome, “how it works,” who it is and is not for, objections, and one call to action. Do not invent customer logos, results, testimonials, or integrations. Include three ad or outreach angles and define what conversion rate or interview count would count as enough evidence to continue. Ask up to 3 clarifying questions only if needed.

Compare these startup options: [IDEAS]. Evaluate them against [DECISION CRITERIA], including problem severity, buyer access, current alternatives, sales cycle, willingness to pay, technical risk, and speed of learning. Create a one-page decision memo with a scoring table, strongest evidence, missing evidence, major disconfirming risk, and recommendation. Do not assume market size, pricing, or demand without support; label those as hypotheses. End with the single cheapest next test for the top two ideas. Ask up to 3 questions only if an input prevents comparison.