- AI Business Insights
- Posts
- The skill was never Excel
The skill was never Excel
I watched someone type the same VLOOKUP four times in one afternoon
In partnership with
It was a Tuesday. Two columns refused to line up, and the fix was to squint at them and try again, and again. One tab over, an AI assistant sat completely idle.
That gap is what caught my eye in a LinkedIn post I came across this week. The argument is simple and slightly uncomfortable: most people are still doing Excel by hand while the thing that could write the formula waits, unused, a keystroke away.
Founders burning hours on lookups. Ops teams flagging duplicate entries one row at a time. Somebody retyping names just to merge two columns. I was nodding along the whole way through, because I have watched people do all three.
Blu Dot surpasses 2,000% ROAS with self-serve CTV ads
Home furniture brand Blu Dot blew up on CTV with help from Roku Ads Manager. Here’s how:
After a test campaign reached 211,000 households and achieved 1,010% ROAS, the brand went all in to promote its annual sales event. It removed age and income constraints to expand reach and shifted budget to custom audiences and retargeting, where intent was strongest.
The results speak for themselves. As Blu Dot increased their investment by 10x, ROAS jumped to 2,308% and more page-view conversions surpassed 50,000.
“For CTV campaigns, Roku has been a top performer,” said Claire Folkestad, Paid Media Strategist, Blu Dot. “Comping to our other platforms, we have seen really strong ROAS… and highly efficient CPMs, lower than any other CTV partner we've worked with.”
Using Roku Ads Manager, the campaign moved from a pilot to a permanent performance engine for the brand.
*Ad
The leverage moved, and most people did not notice
The post frames it as a change in what the job actually is, and I think that is right.
Spreadsheets used to run half the operation. Revenue tracking. Ranking reps. Flagging outliers before they turned into real problems. And back then, an analyst wrote every single formula by hand, because there was no other way to get one.
Now a single prompt does what used to eat a full afternoon. The model does not just answer questions about Excel anymore. It writes the exact formula you need, immediately, in the syntax your version expects.
So the skill is not Excel. The skill is knowing how to ask for it.
That line stuck with me. Nobody is claiming spreadsheets are dead. The claim is that the hard part relocated, and the people who win now are the ones who know what to request.
The twenty prompts, and what each one actually saves you
To make the point practical, the author assembled twenty prompts. I have laid them out with a quick note on what each one buys back, so you can spot the ones that match your own messy files.
Regional revenue totals: add up sales by region in one prompt instead of building nested SUMIFs.
Find and highlight duplicates: surface repeated entries automatically before they skew your numbers.
Merge first and last names: combine two columns into one clean field without retyping anything.
Business days between dates: count working days while excluding weekends and holidays.
Standardize messy text: fix inconsistent capitalization and spacing across a whole column instantly.
Peak sales months per product: spot the strongest month for each item without eyeballing the data.
Letter grades with nested IF logic: assign A through F bands without hand-writing the IF chain yourself.
Split dates into parts: pull the day, month, and year out of any date field.
Clean up formula errors: replace those ugly error cells with clean, usable values.
VLOOKUP or INDEX MATCH across sheets: connect data between tabs without wrestling the syntax.
Winning Startups Aren't Bigger. They're Leaner.
Founders deploying AI agents across sales, marketing, and customer service are closing more with smaller teams. 65% more sales leads. Less headcount.
Download the free Practical Guide to Agentic GTM for Startups and start building the stack your competitors dream of.
*Ad
The second half is where the genuinely tedious work lives
The first ten are the ones people hit weekly. The next ten are the ones that quietly eat a whole morning when they come up.
Self-updating running total: build a cumulative sum that recalculates as new rows come in.
Count keyword mentions: tally how often a word shows up inside longer text strings.
Extract domains from emails: pull the part after the @ from a full list of addresses.
Fix text-stored numbers: convert numbers trapped as text into real numbers you can calculate on.
Compound interest in one formula: run the whole calculation without setting up a manual schedule.
Rank without re-sorting: score data by position while leaving the original order untouched.
Flag outliers by percentage: mark values that fall outside a threshold you set.
Split full addresses: break a single address field into clean street, city, and state columns.
Weighted average across criteria: blend multiple factors into one score with proper weighting.
Pull exact pivot values: grab a specific number straight out of a pivot table with GETPIVOTDATA.
Vague requests get vague formulas
Here is the part I would underline, because it is where most people give up on this and decide the AI is not very good at spreadsheets.
Be specific. Tell it your column letters. Tell it your date format. Tell it what a good result looks like, and what should happen to the rows that do not fit. A prompt like "help me with duplicates" gets you something generic that you will then spend three messages correcting.
A prompt that says "column B holds email addresses with inconsistent capitalization, flag any address that appears more than once, leave the original column untouched" gets you something you can paste straight in.
The formulas were always the hard part. Now the hard part is describing what you want in plain language, and that is a skill anyone picks up in an afternoon.
PRDs by voice. Bug reports by voice. Ship faster.
Dictate acceptance criteria and reproductions inside Cursor or Warp. Wispr Flow auto-tags file names, preserves syntax, and gives you paste-ready text in seconds. 4x faster than typing.
*Ad
The one chore I would hand over first
I like posts like this because they take a boring task and reframe it as an advantage sitting in plain sight. The author is not selling magic. They are pointing at a tool most of us already have open and saying, use it better.
You do not need all twenty on day one. Pick the single chore that annoyed you most this week and let the model write that formula. For most ops people it is the duplicate finder or the messy-text cleaner. For finance it is the weighted average or the compound interest one.
Open the file you have been avoiding. Describe the mess out loud, in a sentence, with your column letters in it. Paste back whatever comes out and see how much of your afternoon you just bought back.