Lesson 07.3: Asking Questions About a Spreadsheet

Module 07: Copilot in Excel Foundations

Lesson 07.3: Asking Questions About a Spreadsheet

Beginner Last verified: 2026-06-02

Mark this lesson complete

Availability note: Copilot in Excel can vary by account, Microsoft 365 subscription, Microsoft 365 Copilot add-on license, Copilot label, app version, workbook format, file location, AutoSave state, tenant settings, admin controls, and rollout status.

Lesson Promise

Ask sharper spreadsheet questions that produce useful insights, not generic commentary.

Real-World Scenario

A founder wants to know what the monthly sales data says, but the workbook contains revenue, refunds, customer segment, channel, and region columns.

Core Concept

A good Excel prompt specifies the business decision, table or range, column names, timeframe, comparison group, and desired output format.

Ask for one analytical lens at a time: trends, outliers, drivers, segments, risks, or questions to investigate. Combining all of them at once usually creates a shallow response.

Copilot can help explain data, but it may miss context that is obvious to the business owner. Add domain context and ask for assumptions to be separated from observed data.

Step-By-Step Workflow

Write the business decision you are trying to support.
Name the table and important columns.
Ask a focused question with timeframe, segment, and comparison criteria.
Request an answer in a table with finding, evidence, assumption, and verification step.
Ask a follow-up question based on the first result.
Capture final conclusions separately from open questions.

Prompt Lab

Bad Prompt

What does this data say?

Better Prompt

Using the SalesData table, compare revenue by channel for Q1 and identify the top three changes. Show columns used and assumptions.

Expert Prompt

Act as a business analyst. Using the SalesData table, answer this decision question: which sales channel deserves more attention next quarter? Compare revenue, refund rate, customer count, and average order value by channel. Separate observed data from assumptions, list possible data quality concerns, and give me three manual checks before I act.

Hands-On Exercise

Ask the same dataset three focused questions: one trend question, one outlier question, and one decision question.

Deliverable

A spreadsheet Q&A log with the question, Copilot answer, columns used, confidence, and manual check.

Excel Review Checklist

Supported workbook format, location, and AutoSave state checked Data formatted as a table or supported range with clean headers Blanks, duplicates, merged cells, subtotals, and data types reviewed Columns used, assumptions, and calculation logic captured At least one formula, filter, chart, or insight verified manually Human review completed before sharing

Common Mistakes

Asking one broad analysis question and accepting the first answer.
Using messy data with duplicate headers, blank columns, hidden assumptions, subtotals, or merged cells.
Treating Copilot's formula, chart, or insight as correct before manual verification.
Mixing private or regulated data into an unapproved workflow.
Promising a feature before checking the learner's license, account, app version, file state, and tenant settings.

Pro tip: Ask Copilot to create a verification companion for every Excel answer: what it used, what it assumed, what could be wrong, and how to check it manually.

Quiz / Checkpoint

What makes an Excel question stronger?

It names the business decision, table, columns, timeframe, comparison criteria, output format, assumptions, and verification method.

Official Sources To Verify

Previous Course hub Next