Month-End Close

Explain income statement flux month over month

Rank income statement movements by size, and get a list of likely causes to check in the general ledger for each one, so your flux review starts with the right questions.

Time
About 20 minutes
Level
Beginner
Works with
Claude, ChatGPT, Microsoft Copilot, Gemini

What you need

  • Income statement by account for the current month, prior month and same month last year
  • Your flux threshold, such as $5,000 and 10%
  • Optional: general ledger detail for the largest movers

Steps

  1. Export the three columns side by side and compute the changes in Excel yourself. Don't rely on AI arithmetic for the numbers.
  2. Paste the prompt and the table.
  3. For each flagged line, pull the general ledger detail the AI suggests and confirm or reject its hypotheses.
  4. Write the final explanations from what the ledger shows, not from the AI's guesses.

The prompt

Fill in the highlighted brackets
Act as a senior accountant preparing an income statement flux review.

Threshold: flag any account where the change versus the prior month OR versus the same month last year exceeds both $[5,000] and [10]%.

Data (account, current month, prior month, same month last year, change amounts):
[paste table]

Company context: [industry, seasonality, anything unusual this month]

For each flagged account, return:
- Account, current amount, $ change and % change, with flagged comparisons marked
- Two or three likely causes, written as hypotheses (for example: "a missed or reversed accrual", "a timing difference in vendor invoicing")
- The specific general ledger detail or source documents I should pull to confirm or reject each hypothesis

Sort by the size of the dollar change. Clearly label all causes as hypotheses to verify, not findings.

Review before you use it

  • All changes were computed in your spreadsheet, not taken from the AI output.
  • Each final explanation is supported by general ledger detail or a source document.
  • Recurring items, like rent, depreciation and insurance, that moved unexpectedly are investigated, not explained away.
  • A reviewer signed off on the explanations.

Never paste into an AI tool

  • Payroll detail by employee
  • Use a business or enterprise AI plan that does not train on your data, and follow your company's AI policy.

Educational content, not professional advice. AI output can be wrong; review it before relying on it. Disclaimer