Turning an Excel Formula Into Text: Three Different Jobs, One Search Term

"Turn this formula into text" is really three different requests wearing one search term, and they call for three different tools. You might want to see a formula written out as a literal string, for documentation. You might want a formula's computed result to stop being a formula at all — frozen as a plain number, so it survives a copy-paste without recalculating. Or you might want a number a formula produced woven into an actual sentence, for a report someone will read rather than a cell someone will audit. Excel has a specific answer to each, and reaching for the wrong one is the most common reason this task feels harder than it should.

Job one: showing a formula as a string

=FORMULATEXT(cell) returns exactly what a cell's formula says, as text, in a neighbouring cell — useful for documenting a workbook, comparing whether two cells that look alike actually contain the same logic, or auditing a sheet someone else built. If B2 contains =SUM(C2:C10), then =FORMULATEXT(B2) in another cell displays the literal string "=SUM(C2:C10)". It returns #N/A if the source cell has no formula at all, which is itself a useful signal — it means whatever produced that value was typed in by hand, not calculated.

Job two: freezing a result as a static value

This is the version most people actually mean by "convert formula to text": stop a cell recalculating and lock in whatever number or string it currently shows. The reliable way is Copy, then Paste Special (Ctrl+Alt+V) and choose Values. That replaces the formula with its last computed result, permanently — the cell now holds a plain number or string with no reference to the cells it used to depend on.

A quicker route for a single cell is F9 while editing it: select the formula in the formula bar, press F9, and Excel replaces the visible formula with its calculated value before you press Enter. Both routes do the same thing; Paste Special scales to a whole range, F9 is faster for one cell you're already editing.

Freezing values matters most right before a sheet leaves your hands — sending it to someone else, archiving a month-end snapshot, or sorting a range where live formulas referencing specific rows would silently point at the wrong data once the sort moves them. A formula that recalculates after a sort is one of the more common ways a spreadsheet quietly corrupts itself, and it looks identical to a correct one until someone checks the numbers against last month's copy.

Job three: wording a number into a sentence

The third job is different again: taking a number a formula produced and turning it into readable prose, usually because it's heading into a report rather than staying in a grid. =TEXT(number, format) does the formatting half — TEXT(A2,"$#,##0.00") turns 1500 into "$1,500.00" as an actual text string, ready to sit inside a sentence built with concatenation or the & operator. ="Revenue was "&TEXT(A2,"$#,##0.00")&" this quarter." produces one clean sentence from a number and a template, no manual retyping required when the number changes.

A worked example

Say B2 holds =SUM(C2:C13), a running annual total that currently evaluates to 84,300. Three different requests, three different results:

• —FORMULATEXT(B2) returns the string "=SUM(C2:C13)" — what the formula is, not what it produced.
• —Copying B2 and pasting values over it turns the cell into a plain 84300, with no formula left at all.
• —="Full-year total: "&TEXT(B2,"$#,##0")&"." produces the string "Full-year total: $84,300." — a sentence built from the number, ready to paste into a memo.

Confusing the second and third is the most common mistake: pasting values gives you a bare number with no formatting or wording attached; building a TEXT string gives you a sentence fragment that is not meant to be summed or charted afterward, because it's text, not a number. Neither is wrong — they answer different questions about what "into text" was supposed to mean.

Where AI actually helps

An AI tool is genuinely useful for drafting the TEXT format string and the surrounding sentence template — getting the punctuation, the number formatting code, and the wording right on the first try is a language task the model is built for, and a specific request with a worked example gets a better answer than a vague one. The research behind that is about language tasks generally, but it holds directly here: show the model one row of real numbers and the exact sentence you want back, and it has something concrete to match rather than a style to guess at.

Asking the same tool to compute the number itself — sum a column, work out the percentage — in a chat window rather than through the actual formula is the wrong use of it. A model predicts plausible-looking text, and a plausible-looking total is not the same as a correctly summed one; this is documented behaviour, not an occasional slip. Keep the arithmetic in Excel's own SUM, and hand the model only the number the formula already produced, to turn into wording.

Let the formula produce the number. Let AI produce the sentence around it. Don't ask either one to do the other's job.

Checks before you trust the result

1. 01Before pasting values over a formula, confirm you actually want to lose the formula — the formula bar shows you exactly what you're about to freeze, and it is not reversible with Undo once the file is saved and reopened.
2. 02If an AI tool drafted a TEXT format string, check it against a number with an edge case — zero, a negative value, a number with more digits than your example — before trusting it on a full column.
3. 03If FORMULATEXT returns #N/A on a cell you expected to hold a formula, that's the tool correctly telling you the value was typed by hand, not a bug in FORMULATEXT.
4. 04Checking an AI answer when you are not the expert applies to a generated sentence as much as to a generated number — read it against the actual figure before it goes into something someone else reads.

What to do Monday

Next time "turn this into text" comes up, name which of the three jobs you actually mean before opening the Paste Special dialog or asking an AI tool for help — it changes the answer completely. If the underlying question is really about a reusable calculation rather than a one-off conversion, an Excel user defined function is the version worth naming and keeping; if it's about comparing group averages rather than wording a single number, that's ANOVA's job, not TEXT's.

If this kind of number-to-sentence conversion runs on a schedule rather than once — a weekly total wired automatically into a template paragraph — treat it the way any automated step deserves to be treated: decide in advance what you would check to notice it silently produced the wrong sentence, in the same spirit as NIST's AI Risk Management Framework for automated decisions generally. A recurring TEXT formula that quietly breaks when a number turns negative is a small version of the same monitoring problem.

For anything heading into an actual written report rather than a single sentence, how to create a report in Excel covers the larger version of this same build-then-verify habit. Both OpenAI and Anthropic publish guidance on giving a model a concrete example rather than an abstract instruction, which is the whole trick behind getting a TEXT sentence right on the first attempt. Coursium teaches this kind of practical, tool-specific judgement directly. Stay ahead of AI by learning the tools on your phone.