Cover Letters Give a Different Signal Than Resumes Do

A resume is a set of claims about you. A cover letter is mostly a set of claims about the reader: this role, this organisation, why the two of you fit. That difference changes the detection question completely. You still cannot tell who typed a letter, and the stylistic tells are as useless here as anywhere. But you can tell whether the letter contains anything that could only have been written after finding out about this specific employer — and unlike authorship, that is a legitimate thing to judge, because it is a judgement about attention rather than about tooling.

This is a reviewer’s article. It is about reading letters, not writing them.

Why the resume tells transfer badly

The signs people list for AI-written text — smooth grammar, no typos, corporate register, an even voice — are worse than useless on a cover letter, because the cover-letter genre demands all four. The convention is formal, polished, and impersonal. A letter that reads exactly like a template is what a careful, anxious applicant produces on purpose, having been told for years that this is the register.

So the base rate defeats you. Nearly every letter in the stack sounds machine-written, including the handwritten ones, and always did.

The signal that is actually there

A letter’s job is to say something the resume cannot. That means it has to make at least one claim that is about this employer and is not recoverable from the job advert alone. Something like: which part of the work you want and why, what you understood about how the team operates, which of your past problems resembles theirs, what you noticed about the product or the organisation that made you apply here rather than somewhere adjacent.

A letter that never does this is not necessarily machine-written. It is a letter written without finding anything out — which is the same result and is the thing you were worried about. The reason letters drafted wholesale by a tool fail this test is not stylistic; it is that the tool had nothing to work from except the advert and the applicant’s resume, so everything in the letter is a recombination of those two documents.

That gives you a concrete read, and it takes one pass:

Cross out every sentence that could appear in a letter to any employer in the same industry. Then cross out every sentence that just restates the advert’s requirements. Read what is left. If nothing is left, the letter told you nothing — and you can say that in a note, because it is an observation about content.

What survives the crossing-out

Three kinds of sentence usually do, and each one is worth reading closely:

A specific reason for this employer. Not “your commitment to innovation” — something particular enough that it would be wrong if pasted into another letter. This is the single hardest thing to fake without effort, because faking it requires the same research as doing it.

A named connection between their problem and the applicant’s experience. “You are moving from one warehouse to three; I did that at a smaller scale and the hard part was the stock reconciliation, not the software.” A drafted letter can produce the shape of this sentence but rarely the content, because the content requires knowing which part was hard.

Something the resume could not say. Why they are changing field, what a gap was, why they want less seniority than they have, what they are trying to move towards. This is the letter’s best use and the reason to read letters at all — it is the place where a nonlinear history gets explained rather than left to inference or where an applicant accounts for something a reviewer would otherwise guess at.

What you are entitled to conclude

Be precise about the boundary, because it is easy to slide across it.

Fair to conclude: this letter contains no employer-specific content; this letter does not explain the career change the resume raises; this letter restates the advert. All of these are properties of the document, all are checkable by a second reader, and all are defensible in a note.

Not fair to conclude: this applicant used AI, this applicant is lazy, this applicant is being dishonest. You do not know any of it. A candidate may have written every word themselves and simply been taught to write letters this way — that is most of the population — and a candidate may have used a tool to phrase a letter whose substance they researched carefully.

The second list is where reviewers get into trouble, and it is avoidable by keeping the note descriptive.

Weight it correctly, then move on

There is a further step reviewers skip: deciding how much the letter should count at all. For many roles the answer is very little. If the letter is optional, if the pipeline receives hundreds of applications, or if the role’s skills are demonstrable in other ways, an uninformative letter is weak evidence about a candidate and a strong signal about your own process — you asked for a document you do not read carefully and they gave you one.

Where the letter genuinely matters — roles that are largely writing, roles where motivation predicts retention, career-change applications where the resume is confusing without it — say so in the advert and ask a specific question rather than requesting a generic letter. “Tell us about a time you had to do X” produces content that cannot be recombined from the advert, and it does so without anyone having to guess at authorship.

The same principle as everywhere else

The durable position is that you are assessing the document’s content, not its provenance. That is the same conclusion the resume version of this question reaches, by a different route: with resumes the substantive test is whether claims are specific and supportable, and with letters it is whether anything in it required finding something out.

And if a letter does raise a suspicion you cannot resolve, the response is the same as for a resume — do not adjudicate it on paper, carry the question to the next stage instead. A letter’s claims about motivation are especially easy to test out loud, because a person who researched an employer has more to say about it than fits in a letter, and a person who did not has less.

One caution for the interview: if the letter and the resume tell slightly different stories, that is worth asking about but is not evidence of anything on its own. Documents written weeks apart drift. Keeping one consistent account across every surface is a discipline, not a default, and the candidates who have not managed it include plenty who wrote both documents themselves.