20 AI Prompts for Excel in Copilot, Claude and ChatGPT

I’ve watched friends type “analyze this” into an AI sidebar in Excel, get back a mushy paragraph, and close the panel for good. I used to do something close to that myself. So when I came across this LinkedIn post from a finance-minded professional who spent years living in spreadsheets, it hit home right away.

The author’s point is simple: most people use AI in Excel like a search bar. They type something vague. They get something vague back. Then they decide AI is overrated.

Meanwhile, they’re still fixing #REF! errors by hand.
Still cleaning the same messy export every Monday.
Still losing an hour to a pivot table.

That’s not a skill. It’s a tax.

The original poster says they paid that tax for years. Budgets, forecasts, scenarios, all of it. Once they started using AI inside Excel, they wished they’d done it sooner. Their takeaway stuck with me: the tool was never the gap. The prompt was. Founders who figure this out early move faster. Same data, same hours, better answers.

The Prompt Upgrade That Changes Everything

Here’s the before-and-after example the creator shared, and it says a lot about the whole idea:

• Instead of: “Analyze this.”
• Try: “Analyze Sheet1 A1:F500, find the top 3 sales decline drivers, and show calculations.”

The second version tells the AI where to look, what to find, and how to prove it. That’s why the answer comes back useful instead of generic. I think most people underestimate how much the AI needs a precise location and a precise goal.

The 6-Part Formula for Strong Excel Prompts

The expert breaks every strong Excel prompt into six pieces. Before you hit enter, check that your prompt covers each one:

1. Context: what the data is about. Tell the AI whether you’re looking at Q3 sales, payroll, or inventory counts.
2. Range: where the data lives. Name the sheet and the cells, like “Sheet1 A1:F500.”
3. Goal: what you need. One clear outcome beats three fuzzy ones.
4. Rules: what must not change. Protect original columns, formulas, or totals.
5. Output: where to show results. A new sheet, a specific column, or a summary in the chat.
6. Check: how to verify it. Ask for calculations, sample rows, or a quick sanity test.

That last one, Check, is the piece I see skipped most often. It’s also what turns AI from a guessing machine into a reliable assistant.

20 Prompts That Cover Real Spreadsheet Work

The person who shared it turned that formula into 20 prompts. They work in Copilot in Excel, Claude in Excel and ChatGPT for Excel. Here’s the full list, with a quick note on why each one earns its spot:

1. Clean messy data into a new sheet. Your original stays safe while the AI tidies a copy.
2. Write a formula and explain it. You get the answer plus the understanding to reuse it.
3. Explain a formula step by step in plain English. Perfect for inherited workbooks with mystery nested IFs.
4. Fix #N/A, #VALUE! and #DIV/0! errors. The classic headaches, diagnosed instead of patched blindly.
5. Look up across sheets, show Not found, flag duplicate keys. A lookup that tells you when something’s off, not just when it works.
6. Build a PivotTable sorted by highest value, with a grand total. No more dragging fields around for an hour.
7. Pick the best chart and say why. The reasoning helps you defend the choice in a meeting.
8. Find trends and outliers, no guessing. The “no guessing” rule keeps the AI anchored to your actual data.
9. Highlight what matters with conditional formatting. Let the important cells jump off the screen.
10. Split and combine text columns. First and last names, addresses, product codes, all handled.
11. Remove duplicates safely, never delete. Flag them first so nothing important vanishes.
12. Build a monthly budget. A clean starting structure in minutes.
13. Forecast with method, assumptions, best, base and worst cases. You see exactly how the numbers were built.
14. Run scenarios on your inputs. Test “what if” questions without rebuilding the model.
15. Build a clean dashboard. One view your team can actually read.
16. Add dropdowns and validation. Stop bad data at the door.
17. Audit a workbook for hardcodes and broken links. Catch the hidden landmines before someone else does.
18. Reconcile two lists. Find what’s missing or mismatched between two sources.
19. Summarize for leadership in 5 bullets. Turn a dense sheet into a message executives will read.
20. Automate a repeated task. That Monday export cleanup? This is where it ends.

If I had to pick a starting point, I’d go with #17. Auditing a workbook for hardcodes is tedious for humans and fast for AI, and it pays off right away.

Habits That Make the Output Better

Good prompts only go so far if your setup is sloppy. The LinkedIn creator shared a handful of habits worth keeping:

• ✔ Use clear table headers
• ✔ Name the exact sheet and range
• ✔ Work on a copy
• ✔ Ask AI to explain before it edits
• ✔ Check formulas on sample rows
• ✔ Never let AI invent numbers

Small habits. Big difference in what comes back. Clear headers alone can turn a confused response into a sharp one, because the AI finally knows what each column means.

The One Rule Above the Rest

This is the part I’d tape to my monitor. The author says to review every output before it touches payroll, taxes, forecasts, financial reports or client deliverables.

AI helps you catch errors earlier. You still own every number you share.

That’s the right mindset. AI speeds you up, but it doesn’t take your name off the report. Treat it like a fast junior analyst whose work you always double-check.

How to Put This to Work Today

If you want to start small, try this sequence:

1. Open a copy of a workbook you use every week.
2. Pick one prompt from the list, like fixing errors or building a pivot.
3. Write it using all six parts: Context, Range, Goal, Rules, Output, Check.
4. Ask the AI to explain its plan before it makes any edits.
5. Spot-check a few rows against the original data.

Do that once and you’ll feel the difference. Do it every week and those lost Monday hours start coming back. I love posts like this because they skip the hype and hand you something you can use in your next spreadsheet session.

Check out the full LinkedIn post to see the original breakdown, and tell the author which prompt you’d run first. You might also want to send it to the teammate who still cleans data by hand.