The short answer
The best AI prompts for founders paste in real material, such as deck text, metrics and customer emails, and ask the model to critique, not cheer. This page has 13 prompts for fundraising, go-to-market, hiring and decisions. Claude is strongest for investor writing and critique (Sonnet 5.5 free, Opus 5.5 on Pro), ChatGPT for metrics and spreadsheets, and Perplexity for market research with citations you can check.
| Prompts on this page | 13, in four groups |
|---|---|
| Best for investor writing | Claude (Sonnet 5.5 free, Opus 5.5 and Research on Pro) |
| Best for metrics and models | ChatGPT (GPT-6 Luna free, GPT-6 Sol on Plus) |
| Best for cited market research | Perplexity (free plan: cited answers, limited daily use) |
| Never hand to AI | Signing decisions, legal and tax advice |
Why most founder prompts give generic advice
Ask an AI assistant "how do I grow my startup?" and you get a blog post. Ask it to rewrite your actual pitch narrative for a seed investor who has heard 40 pitches this week, with your real numbers pasted in, and you get something you can use tomorrow. The difference is context. Every prompt below asks you to paste the real material (your deck text, your metrics, your customer emails) and tells the model what a good answer looks like.
Three habits make these prompts work better:
- Give it the numbers. Revenue, burn, runway, conversion, churn. Without them the model invents plausible ones, which is worse than useless in an investor update.
- Ask for the critic, not the cheerleader. AI assistants lean towards agreeing with you. Prompts such as "argue the investor's side" and "list the three weakest claims" correct for that.
- Verify anything factual. Market sizes, competitor details and legal or tax points must be checked against a primary source or a professional. For research with citations you can open, use a search-first tool such as Perplexity, whose free plan gives cited answers for limited daily use.
Prompts for fundraising and investors
Fundraising is where a sharp outside reader helps most and where errors cost most. The Pressure-test a pitch narrative prompt casts the model as a partner at a specific kind of fund and asks for the questions you will be asked, ranked. The Write a monthly investor update prompt uses a fixed structure that investors skim quickly and forces you to include the bad news and a specific ask. The Prepare for a partner meeting prompt rehearses the hardest questions with short answers you can practise out loud. None of these replace a lawyer for term sheets: use AI to understand the questions to ask, not to decide what to sign.
Prompts for customers and go-to-market
Early go-to-market is mostly writing and listening. The Find the pattern in customer feedback prompt turns a pile of emails, calls and reviews into the reasons people buy, stall and leave, with quotes. The Write a cold email that sounds human prompt keeps outreach short and specific to the recipient, and the Design a pricing test prompt asks for a falsifiable experiment rather than a pricing opinion. For the research behind a new market, see best AI for research and best AI for sales.
Prompts for hiring and team
Founders hire before they have a people team, so the first job descriptions and interview loops are often improvised. The Write a job description that filters prompt focuses on outcomes in the first six months, which attracts people who want the actual job. The Build a structured interview loop prompt produces questions and scoring guidance so every candidate is judged against the same bar. Keep the decision, and anything touching employment law, with people who know your jurisdiction.
Prompts for operations and decisions
The last group is for the founder's own head: a weekly review, a hard decision and a board meeting. The Make a hard decision with a pre-mortem prompt asks the model to imagine the decision failed a year from now and explain why, which surfaces risks that optimism hides. The Draft a board pack outline prompt structures a short pack around decisions, not a tour of every department.
How to adapt these prompts
- Create a company context file. One page covering what you sell, to whom, your stage, key metrics and this quarter's goals. Keep it in a Claude or ChatGPT project so every chat starts informed. Claude's free plan allows up to 5 Projects; ChatGPT sets project limits by plan.
- Swap the reader. Most prompts name an audience (a seed investor, a head of procurement, a senior engineer). Changing that one line changes tone, length and what gets emphasised.
- Ask for options, then choose. "Give me three versions: blunt, warm and formal" is faster than editing one draft three times.
- Run the critic prompt last. Whatever you draft, finish with "List the three weakest sentences and why" before you send it.
Which AI tool works best for founders
For writing investor material, long strategy documents and careful critique, use Claude: Claude Sonnet 5.5 is free, and Claude Pro adds Claude Opus 5.5 and Research. For spreadsheets, quick models and charts from your metrics, ChatGPT is the most flexible; GPT-6 Luna is on Free and GPT-6 Sol on Plus and above. For market research you can cite, use Perplexity. If you want the trade-offs side by side, read ChatGPT vs Claude and Claude vs Perplexity, and for the whole role see AI for founders and best AI for small business.
Your cap table, customer list and financials are sensitive. Before you paste them into a personal account, turn off model training: in ChatGPT it is Settings, Data controls, Improve the model for everyone; in Claude it is Settings, Privacy, Help Improve our AI models. Claude Team and ChatGPT Business do not use your content for training by default.
Prompts to copy
Pressure-test a pitch narrative
You are a partner at a [stage] fund that invests in [sector]. Here is the text of my pitch deck: [paste slide text], and our key numbers: [revenue, growth, burn, runway, retention]. List the ten questions you would ask in the first meeting, ranked by how likely they are to kill the deal. For each, say which slide triggered it and what evidence would answer it.
You get: The investor questions you must answer, ranked
Write a monthly investor update
Write a monthly investor update for [month] using this structure: headline in one sentence, key metrics (vs last month), wins, misses and what we learned, cash and runway, and a specific ask. Facts: [paste metrics, events, hires, problems]. Keep it under 350 words, include the misses plainly, and make the ask concrete (a named intro, a hire, advice on a decision). Do not add numbers I did not give you.
You get: An honest update investors actually read
Prepare for a partner meeting
I have a partner meeting with [fund] on [date]. Their portfolio includes [companies], and their concern so far is [concern]. Give me the eight hardest questions they are likely to ask, a 40-word answer for each using only these facts: [paste facts], and the one question I should ask them.
You get: Rehearsal answers for the meeting that matters
Explain a term sheet in plain English
Explain each clause in this term sheet in plain English: [paste term sheet]. For each one say what it means in practice if the company does well, does badly or is sold, and list the questions I should ask my lawyer. Do not tell me whether to sign; flag anything that looks unusual for a [stage] round so I can raise it.
You get: Plain-English clauses and questions for your lawyer
Find the pattern in customer feedback
Here is raw feedback from [number] customers and prospects: [paste emails, call notes, reviews]. Sort it into three groups: why people buy, why deals stall and why customers leave. Under each, give the top three reasons with how often they appear and one direct quote. End with the single change that would affect the most revenue, and say how confident you are.
You get: Buy, stall and churn reasons with evidence
Write a cold email that sounds human
Write a cold email from me, founder of [company], to [name and role] at [company]. What I know about them: [recent news, their product, mutual connection]. Our offer: [one line] and proof: [customer result]. Under 90 words, one specific reason I am writing to them, no flattery, a low-effort ask such as a 15-minute call. Give me two subject lines.
You get: Short, specific outreach with subject lines
Design a pricing test
We sell [product] to [customer]. Current pricing: [plans and prices]. What we believe: [hypothesis, such as customers will pay more for X]. Design one pricing experiment we can run in four weeks with our traffic of [number] visitors or [number] sales calls a month. Include the change, who sees it, the metric that decides it, the result that would prove us wrong, and the risks to existing customers.
You get: A falsifiable pricing experiment, not an opinion
Write a positioning statement
Help me position [product]. Our best customers are [description], they used to [old way], and they switched because [reason]. Our main alternatives are [competitors or doing nothing]. Write three positioning statements in the form: for [who], who [need], [product] is the [category] that [key benefit], unlike [alternative]. Then tell me which is most believable and why.
You get: Three positioning options and a recommendation
Write a job description that filters
Write a job description for a [role] at our [stage] company. What this person must achieve in the first six months: [outcomes]. Must-have skills: [list]. Honest downsides: [such as ambiguity, small team, travel]. Lead with the outcomes, keep requirements to five, include the downsides plainly, and avoid jargon and words that put off qualified people from applying.
You get: A job post that attracts the right people
Build a structured interview loop
Design an interview loop for [role] with [number] stages. For each stage give the purpose, who runs it, three questions, and what a strong, average and weak answer sounds like. Base it on these outcomes for the role: [outcomes]. Add a scorecard with four criteria scored 1 to 4 so every interviewer judges the same things.
You get: Questions and a shared scorecard for hiring
Weekly founder review
Here is my week: [paste calendar summary, metrics, wins, problems]. My goals this quarter are [goals]. Tell me which activities moved the goals, which did not, what I should stop or delegate next week, and the one thing I am probably avoiding. Be direct and keep it under 200 words.
You get: A blunt weekly review tied to your goals
Make a hard decision with a pre-mortem
I am deciding whether to [decision]. Options: [options]. Context: [numbers, constraints, people involved]. Imagine it is a year from now and I chose [preferred option] and it failed. Write the most likely story of why it failed, list the early warning signs, and suggest what I would need to see in the next 30 days to stay confident.
You get: The risks optimism hides, with early warning signs
Draft a board pack outline
Draft an outline for our [quarter] board meeting. Facts: [metrics against plan, key events, cash position, decisions needed]. Structure it around the two or three decisions I need from the board, then performance against plan, then risks. Keep the pack to [number] pages and suggest one chart for each section.
You get: A decision-led board pack outline