Get one paying customer in six weeks: a pre-build test

The charge is the test

You have an AI-buildable idea, no revenue, and no company yet. Builders using Stripe's payments integration on Replit and Vercel now reach a first customer who pays in under six weeks, with each monthly cohort improving.

Pull out a notebook at the kitchen table. You do not need a deck, a logo, or a company. You need a stranger, a problem, and a charge. For six weeks, try to turn one stranger into a paying customer.

If the charge does not clear, the idea is not ready to build. If it does, you have a small but real reason to incorporate and build the smallest version. Write the deadline in your calendar before you write any code.

iOS app releases rose 24% month-over-month after agentic coding tools that build from prompts became mainstream. More builders means more competition. A charge separates real demand from interest. Speed matters, but speed without a buyer is just activity.

Conversations become commitments

Most founders stop at compliments. A stranger says the problem sounds real, then the conversation fades. You need a written promise: a short note that says they will pay for a specific outcome. Write the outcome, the fee, and the date they expect it.

The count of consumers purchasing AI products increased from fewer than 6 million to 14 million over one year. The signal is broad, not a guarantee. Your job is to find your slice and prove it with a charge.

Stripe's leading AI customer group posted 120% growth in 2025 and 175% growth in 2026. Growth like that does not mean every idea is ready. The buyers are moving, and a founder can move with them.

Do not ask for a free trial. A free trial invites curiosity, not commitment. Ask for a small fee for a concrete outcome. If they hesitate, that is useful information. If they pay, you have a customer, not a fan.

Find strangers where the problem already happens. Look for the places where people describe the pain, ask for workarounds, or compare tools. Do not start with friends. They will be kind. Start with people who have the problem and no reason to be nice to you.

The Six-Week First Charge Test is: 10 conversations, 3 written commitments, 1 paid charge, 1 build/no-build decision. Count a conversation when the stranger names the problem in their own words, a written commitment when they agree to pay for a specific outcome, and the charge only when the payment clears. Make the decision in one short note.

• Ask a stranger to name the problem in their own words.
• Offer a concrete outcome for a small fee.
• Ask for a written promise to pay when it is ready.
• Send a payment link and check for the charge.
• Record the reason for no charge in a short note.

Each item moves the conversation from opinion to proof. Keep the problem in the buyer's words, make the offer concrete, turn interest into a promise, and treat the charge as the only proof. Use the note to decide whether to stop or change.

Build only after the card clears

A cleared charge changes the question. You are no longer asking whether people might pay. You are asking how to deliver the outcome they promised to buy. That is when incorporation becomes useful, not before.

If the charge does not clear, do not build the full product. Change the offer, the price, the buyer, or the channel. Sometimes the idea is real but the packaging is wrong. Sometimes the idea is not ready. The note from the checklist tells you which one you have.

48% of revenue across Stripe's top AI companies came from outside the companies' home markets. If your early conversations all happen inside your local network, you may be testing familiarity, not demand.

At the end of the deadline, write a short decision: build, change, or stop. Build only what the paying customer needs. Change a variable and run the test again. Stop cleanly if the test fails.

Send the payment link now, then write down what you learn.