Skip to content

Use Excel Copilot for Close Analysis and GL Summarization

For Controllers ·

Tool:Microsoft Excel
AI Feature:Copilot
Time:10-15 minutes
Difficulty:Beginner
Microsoft Excel

What This Does

Excel Copilot can summarize general ledger data, build variance analysis formulas, identify unusual transactions, and generate pivot-style summaries from plain-language questions, without manually building pivot tables. For a Controller, this means faster analysis of close data and faster identification of items requiring management attention.

Before You Start

  • You have Microsoft 365 (Business or Enterprise plan with Copilot enabled)
  • Your GL data is in Excel (exported from your ERP system)
  • Data has clear column headers (Account, Amount, Department, Date, etc.). A formatted table helps but is no longer required; Copilot reads plain ranges too
  • You're signed into your Microsoft account

Steps

1. Open Copilot in Excel

With your GL export open, click the Copilot icon in the lower-right corner of the Excel window. Microsoft moved Copilot off the Home ribbon in 2026, so this corner button is now the entry point. Pressing F6 cycles focus to it on any platform, and the newer Alt + C (Windows) and Cmd + Ctrl + I (Mac) shortcuts are rolling out. If you prefer it near the ribbon, right-click the corner button and choose Move to ribbon.

What you should see: A Copilot panel opens on the right side of your screen with a text input box. Troubleshooting: If you do not see the corner button, your Microsoft 365 plan may not include Copilot. Check with your IT administrator.

2. Summarize spending by category

In the Copilot panel, ask a plain-language question about your data:

Type: "Summarize total spending by department for the current month. Rank from highest to lowest."

What you should see: Copilot creates a summary table showing spending by department without you having to build a pivot table manually.

3. Identify large or unusual transactions

Type: "Show me all transactions over $50,000 in the SG&A category this month."

What you should see: A filtered view or new table showing only those transactions, useful for close review and variance explanation.

4. Build a variance formula

For variance analysis, use Copilot to write the formula instead of building it yourself:

Type: "In column G, write a formula that calculates the variance between actual (column D) and budget (column E), and flags 'REVIEW' if the variance is greater than $10,000 or greater than 5% of budget. Otherwise flag 'OK'."

What you should see: Copilot writes the formula, explains what it does, and inserts it in column G.

5. Highlight outliers for review

Type: "Highlight all rows where the amount is more than 2 standard deviations from the average for that account code."

What you should see: Copilot applies conditional formatting or creates a filtered view of statistical outliers, useful for fraud detection and unusual accrual identification.

Real Example

Scenario: It's day 3 of month-end close and you have 2,000 rows of GL detail exported from your ERP. You need to quickly identify what's driving an SG&A overage before the 4pm status call with the CFO.

What you do: Open the GL in Excel, open Copilot, ask: "What are the top 10 largest SG&A transactions this month? Show the vendor/description, amount, and GL account code."

What you get: An instant table of the 10 largest transactions, with no pivot table or SUMIF required. You can walk into the CFO call with the specific vendors driving the overage.

Tips

  • Make sure your GL export has clean, consistent column headers. Copilot needs to know what each column means.
  • Ask follow-up questions in the same session: after getting a summary, ask "Now show me the same summary for last month" for a quick comparison
  • For sensitive variance explanations, always verify the numbers Copilot surfaces against the source ERP. Copilot reads your Excel data, so accuracy depends on data quality.

Tool interfaces change. If a button has moved, look for similar AI/magic/smart options in the same menu area.