Here’s a stat that stopped me mid-scroll: most people use less than 10% of what Excel can actually do. I came across this breakdown from a LinkedIn creator who builds financial models for a living, and I couldn’t stop nodding. If you’ve ever lost an afternoon to a spreadsheet that fought back, this one’s for you.
The original poster shared something that reframed how I think about spreadsheets entirely. According to the author, 63% of advanced users have already stopped doing Excel by hand. They’re not clicking through menus and memorizing formulas anymore. They’re directing the work and letting AI execute it.
What really got me was the honesty in the story. This expert remembers building models by hand for hours: nested formulas, broken references, the whole tangled mess. In the early days of building, that was just a normal Tuesday. Now the same job takes minutes, as long as you know how to prompt it.
Why this shift actually matters
The mind behind this post makes a point I wish more founders understood. The change isn’t about becoming technical. It’s about describing the outcome you want and letting the tool handle the syntax.
No syntax memorization. No manual formatting. Just outcomes.
That’s the part nobody explains early enough. You stop being the person typing formulas and start being the person giving instructions. Big difference.
The tools this expert actually reaches for
One thing I appreciate is that the creator didn’t just hype one magic tool. They matched each tool to a specific job, which is exactly how I like to think about it. Here’s the lineup:
- Copilot: lives inside Excel and edits your live workbook directly. It cleans data, fixes formulas, and explains errors right on the spot.
- ChatGPT: the author’s pick for formula generation and full analysis on messy datasets.
- Claude: the one this professional reaches for when they need VBA explained properly, step by step.
I love that the recommendation isn’t “use one and ignore the rest.” Each tool earns its place. Copilot for live edits, ChatGPT for heavy analysis, Claude for understanding the code behind the scenes.
The step-by-step workflow that’s actually working
This is where the post gets genuinely useful. The creator laid out a sequence you can follow every time you hand a spreadsheet task to AI. I’ve reordered it into clean steps, and each one has a reason behind it.
- Ask the AI to identify your columns first. Before you request anything else, have it map out what each column contains. This keeps every downstream answer accurate instead of guessing at your data structure.
- State the business outcome, not just the formula. Tell it what you’re trying to achieve, like “show me month-over-month revenue growth,” rather than asking for a specific function. The tool picks the right approach when it understands the goal.
- Break big asks into smaller steps. Instead of one giant prompt trying to do everything, split the work. Smaller requests give cleaner, more reliable results and are far easier to troubleshoot.
- Verify every AI formula against known values. Before you trust a formula across thousands of rows, test it on numbers you already know are correct. This catches silent errors before they scale into real problems.
- Never paste sensitive data into a public AI tool. Keep revenue figures and client information out of public tools entirely. This protects you and the people who trust you with their data.
What I like about this order is the logic. You set context, define the goal, work in chunks, confirm accuracy, then protect your data. It’s a full loop, not a random pile of tips.
The payoff the author points to
Here’s the number that stuck with me. The original poster says AI can cut spreadsheet work down by 40 to 60% less time. Think about what that means in practice. A task that ate your whole morning now wraps up before your coffee gets cold.
That’s not a small edge. That’s a founder getting their week back.
I was genuinely impressed reading this, because it lines up with what I keep seeing. The people winning with AI aren’t the most technical ones. They’re the ones who learned to describe what they want clearly and verify the output before trusting it.
My quick take
If you’re still building models by hand, this savvy professional just handed you a shortcut. Start small. Pick one messy spreadsheet you’ve been avoiding, and run it through the workflow above. Identify the columns, state your goal, break it into steps, and check the output against numbers you trust.
The person who shared this framed it perfectly: this is the stuff they wish they’d known sooner. I think a lot of us feel the same way about spreadsheets.
Go read the full LinkedIn post for the complete breakdown, and if someone on your team is still doing this the hard way, send it their way.