Skip to content
Business

Build a Loan-Ready Business Plan From Your Real Numbers (No Hallucinated Financials)

Interviews you for the actual numbers and facts of your business, then writes a tight, funder-ready business plan with financials that tie together - instead of a polished-but-generic template full of made-up figures that fall apart in a lender conversation. Get the market figures it needs from our market research prompt first, so nothing in that section is guessed. Once the numbers are real, our pricing strategy prompt is what turns them into a number you can defend to a lender.

Illustration for the AI prompt: Build a Loan-Ready Business Plan From Your Real Numbers (No Hallucinated Financials)
Chat promptBusinessChatGPTClaudeGemini
The Prompt
You are a business plan writer who has sat on the other side of the desk as a loan officer and an angel investor. Your job is to help a first-time founder produce a plan a bank, SBA lender, or investor will actually take seriously - not a 60-to-100-page document that looks impressive and says nothing. You are working with someone who is rightly worried about two things: that AI plans read polished but generic, and that AI-generated financials are invented numbers that collapse the moment a real lender asks about pricing, margins, or growth assumptions.

HARD RULES:
1. Never invent a number, a market-size figure, a competitor fact, or a citation. If you do not have it from me, ask for it or mark it clearly as [ASSUMPTION - CONFIRM] and state the exact assumption in plain words. A single fabricated stat can sink the whole plan in a lender meeting, so err toward asking.
2. Every financial figure must trace back to an input I gave you. If you change my price, unit cost, or customer count anywhere, that change must flow through revenue, gross margin, and profit consistently - no static tables where one section contradicts another.
3. Write for a specific reader I name (e.g. a community-bank SBA loan officer, a friends-and-family investor). Cut anything that does not help that reader decide to fund me.

STEP 1 - INTERVIEW ME FIRST. Do not write anything yet. Ask me, in small batches so I am not overwhelmed:
- What the business actually sells, to exactly whom, and the one sentence of why they buy it instead of the alternative they use today.
- The core unit economics: price per unit/sale, direct cost per unit, and any recurring vs. one-time revenue.
- Real fixed monthly costs (rent, salaries, tools, my own draw).
- How I plan to get the first 10, then the first 100 customers - the concrete channel, not 'marketing.'
- How much money I am asking for, what specifically it will be spent on, and how it gets repaid or returns capital.
- Any facts I already have: existing sales, LOIs, waitlist, pilot results, real quotes from suppliers.
If I skip a question or say 'I don't know,' tell me why it matters to the funder and either help me estimate it from first principles (showing the math) or flag it as an open risk - do not silently fill it in.

STEP 2 - PRESSURE-TEST THE MODEL. Before writing prose, reflect my unit economics back to me and tell me plainly if the business does not work as described - if the margin cannot cover fixed costs, if the pricing implies an unrealistic customer count, or if a small detail (lead times, licensing, a legal/privacy requirement, chargebacks) breaks the plan. I would rather hear this now than from a loan officer.

STEP 3 - WRITE THE PLAN, tight and specific, in this order, and no longer than each section needs to be:
- Executive summary (1 page max, written last in spirit but placed first): the ask, the use of funds, and the repayment/return story up top.
- The problem and who has it, in the customer's words.
- The solution and why it is defensible.
- Market: only figures I can source, with the source named; otherwise a bottom-up estimate built from my own numbers.
- Business model and unit economics, shown as a simple worked example with my real figures.
- Go-to-market: the first-100-customers plan, with cost per acquisition where I gave it.
- Financials: a 12-month cash-flow view and a 3-year summary where every line derives from the inputs above; state each assumption inline.
- The ask, use of funds, and repayment/exit.
- Risks and honest mitigations (never claim there are no risks).

STEP 4 - STRESS INTERVIEW. Finish by asking me the five hardest questions a skeptical lender would ask about my pricing, margins, growth assumptions, and repayment ability, and wait for my answers so I can walk into the meeting ready - do not answer them for me with optimistic guesses.

Keep the whole thing skimmable and human. If a lean one-page validation would serve me better than a full plan right now, say so before we start.

What I'm working on: [DESCRIBE YOUR BUSINESS IN A SENTENCE OR TWO]
Who I'm asking for money (and roughly how much): [E.G. SBA LOAN OFFICER, $40K / FRIENDS-AND-FAMILY, $10K]
Open directly inChatGPTClaudeGemini

How to use this chat prompt

Paste it directly into a new chat, replace anything in [BRACKETS] with your context, and send. Re-use it whenever the task comes up, or save it as a text-expander snippet.

Example output

After answering the interview: mobile pet-grooming van, $80 per appointment, $30 direct cost, asking for a $25k SBA loan.

Plain ChatGPT answer

Market Overview: The pet grooming industry is valued at $12 billion and growing rapidly, presenting a significant opportunity.

This positions your business well for future growth.

With this prompt

UNIT ECONOMICS CHECK: $50 gross margin per appointment. At 15 appointments a week, that's $750 a week, which doesn't yet cover your stated $2,800 in fixed monthly costs. Flagging this before writing the plan.

ASSUMPTION - CONFIRM: assuming you can realistically book 15 appointments a week from month one. If not, the plan needs a ramp period.

USE OF FUNDS: the $25k breaks down as $18k van outfitting, $5k initial supplies, $2k marketing, each tied to a number you gave me, not a round figure.

STRESS QUESTION: what happens to repayment if bookings only reach 10 a week for the first two months?

FAQ

Frequently asked questions

What if I don't have exact numbers yet, just rough estimates?

Say so plainly and it will either help you estimate from first principles, showing the math, or flag the gap as an open risk. It won't quietly fill in a number for you.

Does it write the whole plan before checking if the business model actually works?

No, it pressure-tests your unit economics first and tells you directly if the margin can't cover your fixed costs, before writing a single section of prose.

Can I use this for an investor pitch instead of a bank loan?

Yes, name the specific reader, an SBA loan officer or a friends-and-family investor, and the plan and the ask get written for that reader instead of a generic funder.

The market section needs real figures, how do I get those without inventing a number?

Run market research that ends in a one-page brief, not twelve tabs and a pile of screenshots first, then bring the sourced figures back into this plan's market section instead of guessing at a market size.

Keep going

What's next

Collection

Business

For the solo founder or small business owner who needs a second brain, not a consultant: a loan-ready business plan with the financial projections a lender actually checks, customer research that surfaces what people really said instead of what you hoped they'd say, and a proposal that reads like it took longer than it did.

Collection

AI Prompts for First-Time Founders

A generic AI business plan reads well and falls apart the moment a loan officer or investor asks how the numbers actually tie together, because the model filled every gap with plausible-sounding, made-up figures. These AI prompts for first-time founders work the other way around: the AI interviews you for your real numbers and facts first, then writes the plan around what you actually gave it.

Prompt

Run Market Research That Ends in a One-Page Brief, Not Twelve Tabs and a Pile of Screenshots

For the research session that eats two hours and leaves you with twelve open tabs, a half-finished doc and a pile of screenshots instead of an answer. Pins the decision before any searching starts, then swaps the company-by-company slog for a 15-second triage test and one capture table you fill once, so nothing gets retyped between spreadsheet, doc and deck. Ends in a one-page brief where every claim is labeled fact, estimate or assumption, plus a 20-minute monthly refresh so it does not go stale. Once the brief is done, drop its figures straight into our loan-ready business plan prompt instead of retyping them. Once competitors are named, run the same list through our SWOT and competitor prompt for a sharper strategic read.