25 AI Prompts for Any Spreadsheet

This spreadsheet is my [sales / customer / expense] data from [time period].
Before we analyze anything:
1) Tell me how many rows and columns it has.
2) Describe each column in plain English.
3) Point out anything that looks messy, missing, or odd. Don't fix anything yet.

In plain English, what does one row of this file represent?
Explain two example rows in a sentence each.

What is the earliest date and the latest date in this data?
Are any weeks or months missing?

I run [type of business]. Based only on the columns in this file, list 10 useful questions this data could answer for me, from most to least valuable.
Say which columns each one uses.

What might I expect this data to tell me that it actually can't, because the information isn't in the file? List what's missing.

Check this file for: duplicate rows, different spellings of the same thing, numbers stored as text, mixed date formats, missing values, and values that look impossible.
List each problem with how many rows it affects and one example.
Don't change anything yet.

In the [column name] column, list every spelling that seems to mean the same thing, grouped together. Suggest one standard name for each group, and wait for my OK before changing anything.

Convert every date in [column name] to the format YYYY-MM-DD.
Dates written like 03/04/2025 are [month/day/year].
Tell me how many dates you changed, and show any you couldn't read.

Apply only these rules: [paste your rules].
Give me the cleaned file as a download, plus a change log table showing each rule and how many rows it affected. If something doesn't fit a rule, flag it instead of fixing it.

List any values that look like typos or mistakes, such as a quantity of 500 or a price of 0. For each one, show the row and say why it looks wrong.
Don't change them. I'll decide.

Which [5] [products] brought in the most revenue from [start date] to [end date]?
Show a table with the name, the revenue, and the percent of the total.
Tell me which columns you used.

Show total revenue for each month as a table.
Which month was best, which was worst, and how big is the gap?

Using [customer ID column], how many customers bought more than once?
How much do repeat customers spend on average compared with one-time customers?

Compare [group A] with [group B] in the [column name] column: number of orders, total revenue, and average order size. Put them side by side in one table, and add one sentence on the biggest difference.

Compare [this period] with [the period before].
What went up, what went down, and by how much? Show the three biggest changes first.

Make a bar chart of [measure] by [category], sorted from highest to lowest.
Label the axis with its units.

Make a line chart of [measure] by [week / month].
Label the highest point and the lowest point with their values.

Make a scatter plot of [column A] against [column B]. Do they seem related?
Describe the pattern in one sentence, and don't claim that one causes the other.

Make this chart ready to share:
- a title that states the main takeaway, using only what the chart shows
- bars sorted from highest to lowest
- every bar light gray except the most important one, in a bold color
- no extra gridlines
- text large enough to read on a phone
Export it as a PNG image.

What is the single most important number in this data for [my goal]?
Give me the number, one sentence on why it matters, and how you calculated it.

How many rows did you use for that answer? Did you leave any out, and why?

Show me exactly how you calculated that: which columns, which filters, and the steps in order.

My spreadsheet says the total is [my number]. You said [the AI's number].
Find the reason for the difference.

Which parts of your last answer come directly from the data, and which parts are your guess? Label each one.

List three ways your answer could be wrong, and what I could check in the spreadsheet to rule each one out.