Business Plan Prompt for ChatGPT, Claude or Gemini (with Examples)

What AI is useful for in a business plan

A business plan is mostly structured thinking: who has the problem, why they'd pay you, how they'll find you and what it takes to get started. An AI assistant is good at turning scattered notes into that structure, asking the questions you skipped and pointing out where your plan relies on guesses.

It is not a source of market data. If you ask for market size, prices or sales forecasts, it may produce numbers that look precise and aren't based on anything. Use it to organize and challenge your thinking, and bring your own figures from real research and real conversations with customers.

What to tell the AI about your business

The quality of the plan depends on the quality of your input. Before asking for anything, write down:

• Your idea in two or three sentences, and what stage you're at (idea, talking to customers, first sales, running business).
• The problem you solve and who has it, as specifically as you can.
• How your customers solve that problem today: alternatives, competitors or do-it-yourself fixes.
• How you plan to make money: one-off sales, subscriptions, services, commissions.
• The channels you'll use to reach customers.
• Your resources: hours per week, startup costs you can cover without putting your personal finances at risk, team and experience.
• What you want to have achieved in 90 days, and what worries you most.

Business plan prompt (lean version)

A short, lean plan is more useful at the start than a 40-page document. This prompt asks for one and makes the AI separate what you told it from what it is assuming.

Act as an experienced small-business advisor. Help me write a lean business plan from the information below.
Idea: [2–3 sentences]. Stage: [idea / validating / first sales / running].
Problem and who has it: [description].
Ideal customer: [who, what they do, where they are].
Current alternatives: [competitors or DIY solutions].
Revenue model: [how I'll charge]. Channels: [how I'll reach customers].
Resources: [hours per week], [startup costs I can cover], [team and experience].
Structure the plan into: problem, customer, value proposition, revenue, costs, channels, competitive advantage and key risks.
Don't invent market data, prices or forecasts. Where a number is needed, leave [TO FILL IN] and tell me how I could find it. Mark every statement that is an assumption I still need to test with [ASSUMPTION].

Based on the plan above, build an assumptions table for me to fill in myself. Columns: assumption, my estimate, how I'll check it, and status (untested / tested).
Include rows for price per sale, expected sales per month, cost per sale, fixed monthly costs, startup costs and cost of getting a customer.
Add three columns for a cautious, a base and an optimistic scenario, all left blank. Explain in one line what each row means; don't fill in any numbers.

Prompts to stress-test your idea

AI tends to agree with you. Ask it explicitly to play the skeptic: the most useful feedback is about what could make the plan fail and what you should check first.

Review my business plan as a skeptical but fair advisor.
1. List the 5 assumptions the plan depends on most, from riskiest to safest.
2. For each one, propose a quick, low-cost way to test it with real customers in the next few weeks.
3. Point out anything vague, contradictory or missing.
4. Tell me the questions a lender or a potential partner would ask that this plan doesn't answer yet.
Be direct and don't soften the criticism.

[paste your plan]

Turn my business plan into a 90-day action plan.
My goal for these 90 days: [e.g. 10 paying customers / validate the price / launch the online shop].
I can work on it [hours] hours a week.
Give me 3 phases of about a month each, with weekly tasks, one measurable checkpoint per phase and what I should decide at the end of each phase if the checkpoint isn't met. Prioritize talking to customers and testing assumptions before spending money.

What AI can't decide for you

• Legal structure, taxes, licenses, permits and contracts: these depend on where you are and your situation. Check your government's official small-business resources and talk to an accountant or lawyer.
• Market figures and forecasts: AI may invent plausible-looking numbers. Use real sources and your own data.
• Whether the business will work: no plan, human or AI, can guarantee results. Testing with real customers is what reduces the risk.
• Funding decisions: before putting in your own money or anyone else's, get advice from a qualified professional.
• Name and brand availability: AI can't check domains, social handles or registered trademarks. Search the official trademark databases yourself.

Keep confidential details (customer lists, contracts, account information) out of the chat. You don't need them to get a useful plan.

Next steps: marketing plan and business name

PromptYa's lean business plan template asks for your idea, stage, problem, customer, alternatives, revenue model, channels, hours, startup capital, team and 90-day goal in a short form. The prompt it generates asks for a lean plan, hypotheses to validate, a 90-day plan and a financial assumptions table that you fill in yourself.

When you need a name, the name brainstorm template proposes names by family, with slogans and a checklist of what to verify. You can copy each prompt or open it directly in ChatGPT, Claude or Gemini.