How to Turn an Unpaid Invoice Spreadsheet Into a Daily Collections Worklist With AI

The problem and who this is for

A long unpaid-invoice list is not a worklist. It is just a pile. If you open your accounting export each morning and still do not know who to contact first, who needs a phone call, and who can wait, you need a better triage view.

This workflow is for small business owners, bookkeepers, and office leads who already have an accounts receivable export and want a practical daily collections list.

Prerequisites

You need a spreadsheet export with one row per invoice. At minimum, include customer name, invoice number, invoice date, due date, balance due, and any last-contact or status field you already track.

Use ChatGPT as the primary path when you want spreadsheet analysis and quick prioritization. Gemini is a good fallback if your file already lives in Google Drive or you prefer Gemini's file workflow.

How to capture or gather the source material

Export a fresh unpaid-invoice file from your accounting system. Before uploading it, clean it enough that the model can read it:

• one row per invoice
• one header row
• no merged cells
• no subtotal rows
• no blank decorative columns
• no extra tabs unless they are necessary

If you cannot export cleanly, copy the unpaid invoice data into a simple CSV with only the columns you need.

Step-by-step workflow

1. Shrink the file to the working view you actually need.
   Do not upload your entire accounting database export. Start with open invoices only.

2. Define the priorities before you ask for a worklist.
   Decide what matters most. Common factors are age past due, amount owed, customer importance, payment history, and whether anyone has contacted the customer recently.

3. Ask the model to score, not just summarize.
   A useful output is not "these invoices are overdue." A useful output is a scored list with next actions.

4. Generate a daily collections table.
   Ask for columns such as priority level, reason for priority, next action, draft follow-up style, and owner.

5. Sanity-check the top ten rows.
   AI can spot patterns, but you should still review the top of the list before acting on it.

Tool-specific instructions

Primary workflow: ChatGPT for spreadsheet triage

ChatGPT's data analysis workflow is a good fit when you want the file turned into a sortable table, a quick explanation of why rows were prioritized, and optional charting or summary counts.

Fallback: Gemini for uploaded spreadsheet review

Gemini can analyze uploaded spreadsheets and other files. This is useful if your file is already in Google Drive or if you want to keep your work inside the Google ecosystem.

Copy/paste prompt block

{
  "task": "Turn an unpaid invoice spreadsheet into a daily collections worklist",
  "role": "You are helping a small business prioritize collections work for today.",
  "context": {
    "goal": "Create a ranked worklist from the uploaded unpaid invoice file.",
    "priority_factors": [
      "days past due",
      "balance due",
      "customer importance",
      "last contact date",
      "open dispute flag"
    ],
    "business_rules": {
      "small_balance_threshold": "[Insert amount or write none]",
      "high_value_customer_definition": "[Insert definition]",
      "escalation_rule": "[Insert rule or write none]"
    }
  },
  "instructions": [
    "Read the uploaded spreadsheet and assume each row is one open invoice unless the file clearly shows otherwise.",
    "Rank the invoices from highest to lowest priority for today's follow-up.",
    "Create a table with these columns: customer_name, invoice_number, balance_due, days_past_due, priority_level, reason_for_priority, next_action, suggested_message_type.",
    "Use plain English for the reason and next action.",
    "Do not invent facts that are not in the file.",
    "If the file is messy or missing fields, explain what is missing before ranking."
  ],
  "output_format": {
    "deliverables": [
      "ranked_worklist_table",
      "top_10_summary",
      "obvious_data_issues"
    ]
  }
}

Quality checks

Review the first ten rows and confirm that the ranking makes business sense. If a two-day overdue invoice outranks a ninety-day overdue invoice, there should be a real reason. Also check whether any disputed invoice was ranked as a standard payment chase. That usually means the model missed a status field.

Common failure modes and fixes

Failure mode: The spreadsheet is too messy for a good ranking.
Fix: Strip the file down to open invoices only and re-upload.

Failure mode: The priorities feel arbitrary.
Fix: Add your own ranking rules before the prompt, such as "days past due matters more than invoice size."\n\nFailure mode: The output is a summary instead of a worklist.
Fix: Explicitly ask for a row-by-row table with next action.

Failure mode: The model ignores your last-contact field.
Fix: Name that column directly in the prompt and tell the model to penalize recently contacted accounts.

Sources Checked

• OpenAI Help Center, "Data analysis with ChatGPT." Accessed 2026-03-19. https://help.openai.com/en/articles/8437071-data-analysis-with-chatgpt
• OpenAI Help Center, "File Uploads FAQ." Accessed 2026-03-19. https://help.openai.com/en/articles/8555545-file-uploads-faq
• Google Gemini Apps Help, "Upload & analyze files in Gemini Apps." Accessed 2026-03-19. https://support.google.com/gemini/answer/14903178

Quarterly Refresh Flag

Review this article by 2026-06-17 to confirm supported spreadsheet file workflows and limits are still current.