Analyse Your Sales Data With AI and Excel in 30 Minutes

You have a spreadsheet of sales from your billing software, marketplace or accounts system. You know there are useful answers inside: which products are growing, which customers stopped buying, which month was weak. But where do you start?

This walkthrough shows how to go from raw export to three clear findings in about 30 minutes, using Excel for the numbers and an AI assistant as your guide.

The golden rule

Excel calculates, AI guides. Let the AI suggest questions, formulas and pivot layouts, but do the actual calculations in Excel on your full data. AI assistants can make arithmetic mistakes, and pasting thousands of rows of customer data into a public chat tool is not a good idea anyway.

Step 1: Export and look at the data (5 minutes)

Export sales for the last 6 to 12 months as Excel or CSV. Ideally one row per invoice line, with columns like:

Then describe the structure to the AI, with a few rows of dummy data (replace real customer names):

I have a sales export in Excel with these columns: [list]. About 6,000
rows, April to October 2026. Here are 3 sample rows with dummy names:
[paste]. What problems should I check for before analysing, and how?

Step 2: Clean the data (5 to 10 minutes)

Typical problems and quick fixes:

• Title rows or totals inside the data: delete them.
• Dates stored as text: convert them to real dates (the AI can give you a formula).
• Extra spaces in names: use TRIM in a helper column.
• Returns or cancelled invoices: decide whether to exclude them and mark them clearly.

Then select the data and press Ctrl+T to make it a table. Check that the total of the Amount column matches the total in your source system.

Step 3: Ask the AI what to look at (2 minutes)

I run [type of business]. With this sales data, what are the 6 most
useful questions I should answer to make better decisions next quarter?
For each, tell me whether to use a pivot table or a formula in Excel.

Typical questions:

1. How are total sales trending month by month?
2. Which products or categories are growing or shrinking?
3. Who are the top 10 customers, and how much of total sales do they make up?
4. Which regular customers have not bought in the last 60 days?
5. Which cities or regions perform best and worst?
6. What is the average invoice value, and is it changing?

Step 4: Build the summaries (10 minutes)

Use Insert → PivotTable for most questions. Ask the AI for the exact layout when you are unsure:

For "sales by category by month", what should I put in Rows, Columns and
Values in the pivot table, and how do I group dates by month?

For "customers who stopped buying", a formula approach works well. Ask:

In my table Sales, how do I list each customer with their last purchase
date and total sales, and flag customers whose last purchase is more than
60 days ago? Excel 365.

Check every summary: the grand total of each pivot should match the total of the Amount column.

Step 5: Turn numbers into findings (5 minutes)

Copy your summary tables (not the raw data) into the AI:

Here are my summary tables:
[paste monthly sales, category trend, top customers]
Write 3 key findings for the business owner, each with the supporting
numbers, and one suggested action per finding. Use only these numbers.

Then verify every number in the findings against your pivots. Remove any finding that the numbers do not clearly support.

Good findings are specific and actionable, for example:

• "Category B grew from ₹2.1 lakh to ₹3.4 lakh a month since July. Increase stock before the festive season."
• "8 of our top 30 customers have not ordered in over 60 days. Call them this week."
• "Average invoice value fell 12% since August. Check whether discounts are being overused."

Step 6: Make it repeatable

Save the workbook as a template. Next month, paste the new export into the same table, click Data → Refresh All, and your pivots update. Ask the AI for fresh findings from the new summaries.

Common mistakes to avoid

• Trusting a total the AI calculated instead of Excel
• Analysing data that still contains duplicate or cancelled invoices
• Reading too much into one month; look for patterns over several months
• Pasting full customer lists into a public AI tool

Learn more

Pivot tables, lookups, data cleaning and one-page reports are covered step by step in our free course AI for Excel: Beginner to Confident.