As of 13 August 2026, AI can identify your unpaid invoices.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open your invoicing or accounting system and export all invoices for the period into CSV or Excel, including invoice number, customer name, invoice date, due date, total, credit notes and current status.
   2. Open your business bank account or payment processor and export transactions covering the same period and a reasonable payment window after the latest invoice, including date, amount, payer or reference and transaction type.
   3. Remove unrelated transactions from both files, preserve the original invoice numbers and references, and save copies of the untouched exports before uploading anything.
   4. Paste or upload both files into an AI chat with the prompt, filling in the check date and identifying which file is the invoice export and which is the payment export.
   5. Open the AI's unpaid and exceptions sections and compare every item against the invoice record and the corresponding bank transaction, correcting matches where the payment reference or customer name is ambiguous.
   6. Update your accounting or invoicing system only after the checked list agrees with the bank and ledger, then send payment reminders only for invoices that remain unpaid after those checks.

   Prompt

   I need to identify unpaid invoices for a UK business. I will provide two current files: an invoice export and a bank or payment export. Treat the invoice export as the source for invoice numbers, customer names, invoice dates, due dates and amounts, and treat the payment export as the source for payments received. Do not invent or estimate any values.
   
   Invoice data:
   [PASTE OR UPLOAD INVOICE EXPORT]
   
   Payment data:
   [PASTE OR UPLOAD BANK OR PAYMENT EXPORT]
   
   Match payments to invoices using invoice number or payment reference first, then customer name, amount and date. Allow for partial payments, overpayments, duplicated rows, credit notes, refunds, bank fees and payments received under a different name. Do not mark an invoice as unpaid solely because a reference is missing. Produce four sections:
   1. Matched and fully paid invoices.
   2. Part-paid invoices, showing the invoice amount, payments matched and balance remaining.
   3. Invoices that appear unpaid, showing invoice number, customer, due date, amount, days overdue as at [CHECK DATE], and the evidence for the conclusion.
   4. Exceptions requiring manual checking, including ambiguous matches, missing records, credit notes, duplicates and payments that could not be allocated.
   
   For every invoice in the unpaid section, state the exact row or rows used and label the result as confirmed by the supplied data or needs manual confirmation. Do not draft payment reminders and do not give legal or debt-collection advice. If the files do not contain enough information for a reliable match, say what is missing.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.