Build a Custom GPT That Writes Overdue Invoice Reminders

Ask plain ChatGPT for an overdue invoice reminder and you get all caps, emojis and threats of legal action. In this free lesson you build a payment reminder writer for Maple Street Bookkeeping: a Custom GPT that follows the firm's collections policy and writes a reminder that is polite, firm and matched to how late the invoice is.

What you will build: a payment reminder Custom GPT

You start by asking ordinary ChatGPT one bare question: write a payment reminder to a client whose $850 invoice is 45 days overdue. The reply is the email nobody should send. The subject line says URGENT and FINAL NOTICE. It calls the client a Valued Customer, says the firm is very disappointed, and threatens fees, collections and legal action within 48 hours. There is no invoice number, no due date and no way to pay.

Then you build a Custom GPT called Payment Reminder Writer. You give it written rules and a collections policy document, and send the exact same prompt again. This time it writes a real reminder: a subject line with invoice #1042, a greeting by first name, the amount, the original due date and the days overdue, three ways to pay, a request for payment within 7 days, and an offer of a short call or a payment plan. It reads like someone who wants to get paid and keep the client.

Who this is for

This lesson is for the people who chase money at a small business: the bookkeeper, the office manager, the owner who does the books on Sunday night. Maybe you work at a small bookkeeping or accounting firm and send reminders for your own invoices. Maybe you handle accounts receivable for a shop or a service business. Either way, writing the fourth reminder of the week is draining, and it is easy to sound too soft or too harsh. You do not need any AI experience. If you can copy and paste, you can do this.

Step by step: what you do in the lesson

• Ask ChatGPT: click the message box, paste the bare reminder prompt, send it and read the reply.
• Explore GPTs: open More in the sidebar, click GPTs, then + Create to open the GPT builder.
• Name your GPT: call it Payment Reminder Writer.
• Add instructions: paste the reminder rules for Maple Street Bookkeeping. They set the tone, how firm to be at each stage, what every email includes and what it never says.
• Upload knowledge: open Google Drive, preview Maple-Street-Collections-Policy.pdf, scroll through it and click Select to attach it.
• Test your GPT: paste the same prompt word for word into the preview panel, send it and compare the two answers.

What goes into the reminder rules and the collections policy

The instructions read like a brief for a new hire. Sound calm and friendly, never like a debt collector. No emojis, no all caps, no exclamation points. No threats, guilt or shame, and never the words final warning or legal action. Every email has a subject line with the invoice number, a first-name greeting, the invoice number, amount, due date and days overdue, the ways to pay and one next step, all in under 150 words.

The core of it is tone by days overdue. From 1 to 15 days it is a friendly note that assumes the client simply missed it. From 16 to 30 days it is a clear reminder asking for payment or a date. From 31 to 60 days it is polite and firm: pay within 7 days, and here is a call or a payment plan if timing is hard. At 61 days or more the GPT does not write the email at all. The owner calls the client personally.

The collections policy PDF repeats those stages and adds the payment terms (due 30 days after the invoice date; pay by link, bank transfer or check), a rule to only mention a late fee if the engagement letter includes one, a rule for clients who say they already paid, and a sample reminder. The numbers and terms in the lesson are examples. Use your own policy.

Why it works

Plain ChatGPT does not know your terms, your clients or how you like to handle late payers, so it guesses, and the guess is loud. A Custom GPT carries your rules and your policy into every chat. You stop re-explaining the firm each time, and every reminder comes out with the same facts in the same order. The stage rules also stop the two mistakes that cost the most: being so soft the client does not realize it is late, and being so harsh the client leaves.

How to use this in your bookkeeping or finance work

Swap in your own collections policy, payment options and a reminder you were happy with. Then, when you run your accounts receivable aging report, give the GPT the client's first name, invoice number, amount, due date and days overdue, and let it draft the email while you check the numbers. It fits well into a weekly AR routine. Keep a human in the loop: you read and send every reminder yourself, and you decide what happens with anything past 60 days. This lesson is about clear, respectful writing. It is not legal advice about collections.

Frequently asked questions

Can ChatGPT write payment reminder emails for overdue invoices?

Yes, but a bare prompt tends to give a generic or aggressive email. A Custom GPT with your collections policy and a few rules writes reminders that include the right facts and match the tone to how late the invoice is.

Is this lesson free?

Yes, Stepthrough is completely free. To build the GPT for real you need your own ChatGPT account, and creating Custom GPTs may require a paid ChatGPT plan.

Will the GPT send emails to my clients on its own?

No. In this lesson the GPT drafts the email and you copy it into your own email. You review and send every reminder yourself.

Does this give legal advice about debt collection?

No. The lesson teaches you to write polite, clear reminders based on your own policy. The fees and terms shown are made-up examples. For questions about late fees or collection rules, talk to a qualified professional.

How does the interactive lesson work?

Each step is shown to you first on the left (Watch). Then you do it yourself in a practice copy on the right (Your turn), and every click is checked. If you get stuck, press Do it for me. It is a practice copy, so no real accounts are touched.

More in AI Automation for Finance