I used to lose whole afternoons to one stubborn VLOOKUP. You know how it goes: the wrong column index, a #N/A staring back at you, and a cold coffee next to the keyboard. So when I found this post from a LinkedIn creator who has clearly watched this happen again and again, it hit home.
The author makes a simple point. Most people are still doing Excel by hand while AI sits right there, unused. Founders spend hours on VLOOKUPs. Ops teams flag duplicate entries one row at a time. Someone retypes names just to merge two columns. None of it needs to happen anymore.
Why this shift matters
As the original poster explains, spreadsheets used to run half the operation. They handled revenue tracking and rep rankings, and they flagged outliers before those turned into real problems. Back then, analysts wrote every formula by hand. Now a single prompt does what used to take a full afternoon.
AI doesn’t just answer questions anymore. It writes the exact formula you need, instantly.
According to the expert, most people haven’t caught up to that shift yet. So they put together 20 prompts that turn Excel into an AI-powered tool. I’ve added a short note to each one so you know when to use it and which formula the AI will probably hand back.
The 20 prompts, explained
- Calculate regional revenue totals in one prompt. Ask for a formula that adds up sales by region. You’ll usually get SUMIFS, which gives you a clean summary without building a pivot table.
- Find and highlight duplicate entries automatically. The AI can write a COUNTIF rule for conditional formatting, so any value that shows up more than once lights up. That’s great for customer lists and invoice IDs.
- Merge first and last names into one column. Instead of retyping, ask for a formula that joins two cells with a space. It’s a tiny task that eats a surprising amount of time when you have 2,000 rows.
- Calculate business days, excluding weekends and holidays. NETWORKDAYS handles this, and the AI will show you how to point it at a holiday list. It’s perfect for project timelines and SLA tracking.
- Standardize messy text formatting instantly. Extra spaces, random capitals and hidden characters disappear when you combine TRIM, PROPER and CLEAN. Describe the mess and let the AI build the fix.
- Find peak sales months per product. Ask which month had the highest sales for each product. Expect a mix of MAXIFS and INDEX MATCH that does the hunting for you.
- Assign letter grades with nested IF logic. Nested IFs are where most people give up. Describe your score bands and the AI writes the whole chain, or a cleaner IFS version.
- Extract day, month, year from any date. DAY, MONTH and YEAR split a date into parts you can filter and group. That’s handy for seasonal reports.
- Replace formula errors with clean values. IFERROR swaps ugly #DIV/0! or #N/A results for a zero, a blank or a friendly note. Your dashboards instantly look more professional.
- Run VLOOKUP or INDEX MATCH across sheets. Tell the AI where your lookup value lives and which sheet holds the answer. It handles the sheet references that usually trip people up.
- Build a running total that updates itself. A SUM with one anchored cell reference grows as you drag it down. New rows keep the total current without any extra work.
- Count keyword mentions inside longer text strings. COUNTIF with wildcards finds a word buried inside comments, reviews or support tickets. It’s a quick way to spot trends in feedback.
- Extract domains from a list of emails. Ask for everything after the @ sign. That’s useful for grouping leads by company or spotting duplicate accounts.
- Convert text-stored numbers into real numbers. Numbers stuck as text won’t add up, and the error is easy to miss. VALUE or a quick multiply-by-one trick fixes the whole column.
- Calculate compound interest in one formula. Give the AI your principal, rate, compounding periods and years. It returns a formula you can reuse for savings, loans or growth projections.
- Rank data without touching the sort order. RANK.EQ scores each rep or product in place. Your original layout stays exactly where it is.
- Flag outliers based on percentage thresholds. Ask for a formula that marks any value more than, say, 20 percent above or below the average. This is the early warning system the author mentioned.
- Split full addresses into clean columns. Street, city, state and ZIP go into separate cells with TEXTSPLIT or a combination of text functions. Mailing lists and CRM imports get much easier.
- Calculate a weighted average across criteria. SUMPRODUCT divided by the sum of the weights gives you a fair score. It works well for vendor comparisons, grading and lead scoring.
- Pull exact values from a pivot table with GETPIVOTDATA. This one scares a lot of people. The AI explains the syntax and builds the formula, so your reports reference pivot data reliably.
How to get great formulas every time
The creator sums it up well: none of this is complicated once you know what to ask for. In my experience, the quality of the formula depends on how much context you give. Here’s a simple structure that works in ChatGPT, Claude, Copilot or Gemini:
- Describe your layout. Say which columns hold which data and where the rows start and end.
- State the goal. Say what result you want and where it should appear.
- Ask for an explanation. A one-line breakdown helps you learn the logic and catch mistakes.
For example, you might type: “Write an Excel formula that sums column C when column A matches the region in cell F2. My data runs from row 2 to row 500. Explain how it works in one sentence.” That’s specific enough to get a working answer on the first try.
One more tip: always test the formula on a few rows you can check by hand. AI is fast, but you’re still the one signing off on the numbers.
The real skill now
This is the part I keep thinking about. As the post’s author puts it:
The skill isn’t Excel anymore. It’s knowing how to prompt it.
That changes who gets to be the spreadsheet person on the team. You don’t need to memorize hundreds of functions. You need to describe your problem clearly. That’s a much easier skill to build, and it carries over to every other tool you use.
The original poster suggests sending this to the teammate who’s still doing it the slow way, and I agree. Pick one prompt from the list, try it on a real file this week and watch how much time comes back to you. Which one are you trying first?
Check out the full LinkedIn post for the complete list and the author’s take on why prompting is the new spreadsheet skill.