Create a copy-paste prompt library for a startup at this stage: [STARTUP STAGE]. The company is [STARTUP DESCRIPTION], serving [TARGET CUSTOMER], with this current evidence and traction: [EVIDENCE]. Our immediate business objective is [OBJECTIVE], and our constraints are [CONSTRAINTS]. Produce 10 prompts, selected from customer discovery, problem definition, market mapping, positioning, pricing research, landing-page copy, onboarding, experiment design, founder sales, operating metrics, hiring, investor updates, and board preparation. Choose only prompts that fit the stage and objective; do not pad the list with generic brainstorming. For each prompt, provide: a task title, when to use it, the complete copy-paste prompt, and a one-sentence note on the evidence to supply. Each prompt must contain useful bracketed inputs, specify a concrete deliverable, and force the model to separate facts, assumptions, and open questions. For customer research, require interview questions that avoid leading respondents. For market or competitor work, require sources and uncertainty labels rather than invented facts. For investor material, prohibit fabricated traction, market size, customer quotes, or forecasts.
Organize the library in priority order for the next 30 days. End with a two-week sequence showing which three prompts to run first and what decision each should inform. Before answering, check that each prompt has a distinct decision or deliverable and that no prompt asks the model to invent evidence. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [STARTUP STAGE] | Enter the stage, such as pre-idea, validation, pre-seed, seed, or growth. | pre-seed with a working pilot |
| [STARTUP DESCRIPTION] | Enter a short description of the product, buyer, and business model if known. | software that helps independent dental practices recover unscheduled treatment plans |
| [TARGET CUSTOMER] | Enter the primary user or buyer and their context. | practice owners and office managers at 1-5 location dental groups |
| [EVIDENCE] | Enter current interviews, revenue, product usage, or other evidence, including gaps. | 18 interviews; 3 unpaid pilots; users report staff follow-up is inconsistent; no revenue yet |
| [OBJECTIVE] | Enter the most important outcome for the next 30 days. | decide whether to charge for pilots and identify the strongest buyer message |
| [CONSTRAINTS] | Enter limits on time, team, budget, geography, regulation, or available data. | two founders, $8,000 remaining budget, no access to patient health data |
How to use
- State the stage and evidence plainly; weak evidence is more useful than polished assumptions.
- Generate the library, then copy only the first prompt tied to the decision you need now.
- For research outputs, require links or source names and mark unsupported conclusions as hypotheses.
- Follow up with: "Turn prompt 2 into a 30-minute interview guide for office managers, with a neutral opening and no leading questions."
Variations
Customer discovery
Use this before committing to a problem or solution.
I am testing this startup idea: [IDEA] for [TARGET CUSTOMER]. What I know so far is [CURRENT EVIDENCE]. Create a customer-discovery research pack: recruiting criteria, a neutral outreach note, a 30-minute interview guide, follow-up probes, and a synthesis table. Focus on current behavior, workarounds, urgency, budget ownership, and switching triggers. Do not ask respondents to validate my solution. Label hypotheses separately from evidence and flag claims that require more interviews. Ask up to 3 questions only if a required input is missing.
Investor update
Use this to write a factual monthly investor note.
Draft a concise monthly investor update for [COMPANY] using only these facts: [MONTHLY FACTS]. The recipients are [INVESTOR AUDIENCE], and our one request is [ASK]. Output an email with headline, progress, metrics with period comparisons where supplied, learnings, risks, next-month priorities, and the ask. Explain misses directly without spin. Do not manufacture benchmarks, customer quotes, pipeline, or projections. Check that every number and claim appears in my facts, then flag any ambiguous metric definition.
Experiment brief
Use this when a team needs to test one growth or product assumption.
Write an experiment brief for [HYPOTHESIS] affecting [USER SEGMENT]. Current baseline is [BASELINE], proposed change is [CHANGE], and limits are [CONSTRAINTS]. Output the decision being tested, hypothesis, primary metric, guardrail metrics, audience and sample assumptions, implementation steps, duration, pass/fail threshold, and interpretation rules. Include likely confounders and a no-result outcome. Do not claim statistical confidence without sample and variance information. Check that the metric can actually change because of the proposed intervention.
Tips
- Use prompts to prepare a decision, not to generate a startup strategy in the abstract. Name the decision, deadline, evidence, and owner before you paste one.
- Keep raw interview notes separate from conclusions. Ask the model to tag direct quotes, observed behavior, and interpretation so an attractive narrative does not outrun the evidence.
- Do not use AI-generated market-size figures in an investor deck without finding and verifying the underlying source.
- For early-stage positioning, test language in customer conversations and landing-page behavior before treating copy quality as product-market fit.
FAQ
Can AI validate a startup idea?
AI can help structure research and expose assumptions, but it cannot replace interviews, observed behavior, or paid commitments from the target customer.
What startup task should I prompt first?
Start with the decision blocking the next two weeks, often a customer interview plan or a synthesis of existing evidence.
Can I use these prompts for investor materials?
Yes, for structure and editing. Supply verified facts and do not let the model fill gaps with invented metrics or market claims.