Build a customer receivables ageing table

A draft row-level ageing table, totals by customer and currency, and a missing-date/dispute exception list.

Choose one of three approaches. The steps for that option appear below.

Choose an option above. On a phone, swipe the tabs left or right to see all options →

ChatGPT calculates the draft; you check the code

ChatGPT runs a code-backed batch calculation on the approved export. You inspect the code and recalculate samples before keeping anything.

  1. You

    Prepare the inputs

  2. ChatGPT

    Produce the draft

  3. You

    Check and save

  1. Make a local draft from authorised inputs

    Use: You

    Have ready: customer open-items CSV (columns listed in the prompt) and owner-approved rules.

    Check

    Stop if upload permission is absent. Verify date, currency, completeness and the supplied rules; do not replace missing rules with guesses.

    Details

    Inputs to have ready

    Inputs needed: a customer open-items CSV with item ID, customer ID, invoice date, due date, currency, open amount, credit-note link and dispute status (or still to confirm); an owner-approved as-of date, date convention, ageing bucket boundaries and credit-note treatment.

  2. Produce the draft table and show the code

    Use: ChatGPT

    Ask ChatGPT's data analysis for the row-level table, totals by customer and currency, and the exception list, with Python code and row counts.

    Details

    What to paste into the tool

    Task: Produce a draft row-level ageing table, totals by customer and currency, and a missing-date/dispute exception list. Inputs: [ATTACH: the approved open-items CSV from step A-1] [PASTE: as-of date] [PASTE: date convention, including input date format and how days past due are counted] [PASTE: ageing bucket boundaries, e.g. 1-30, 31-60 days past due] [PASTE: credit-note treatment]. Use only these inputs. Show the Python used and a row-level audit trail with source IDs. Age by days past due, not invoice age. Keep not-yet-due balances separate. Apply only the supplied bucket boundaries, date convention and credit-note treatment. Keep disputed items flagged and currencies separate. Missing due dates go to the still-to-confirm bucket, not the oldest bucket. Never silently drop or repair rows. Report missing facts as not stated. Do not access accounts or external sources, send messages, post entries, change the master ledger, sign, publish or pay. Produce a draft for human review only.

  3. Check, correct and save the draft pack

    Use: You

    Check the output against the originals and correct or reject it; save only the checked pack locally, marked draft.

    Check

    Recalculate days past due at boundaries, negative balances and blank due dates. Check every input item appears once. Totals match open amounts by currency. Do not send, sign, pay, publish, post journals or update master records. No accounting-system action.

Before you paste into any AI tool, remove any confidential information. Put it back at the end.

Tools in this workflow

Deciding between ChatGPT and Claude for this? See the comparison.

Final check

Never paste credentials or full payment details.

Confirm authorised inputs and rules before use.

Recalculate days past due and bucket assignment for the as-of date, due today, one day overdue, each boundary, negative balances and blank due dates. Tie every currency total back to the source export.

Reconcile completeness and totals, preserve open questions, and save only a draft.

Check the result against your original inputs. Approve it before sharing or acting.

Availability notes

Based on official product documentation checked on 2026-10-29.

What still needs checking
  • Results on your own data depend on your inputs and rules. Always check the output yourself.
  • Owner-approved inputs and policy conventions must be supplied at use time.

Share this page