ChatGPT can get you from a blank page to a working draft in under a minute. The problem isn’t speed — it’s that a cover letter generated from a generic prompt reads like it was generated from a generic prompt, and hiring managers who read dozens of these a week can spot the pattern immediately.
The fix isn’t avoiding AI. It’s prompting it with the same specific detail a strong cover letter needs anyway, then editing out what gives it away.
Feed it specifics, not a job title
“Write me a cover letter for a marketing manager position” produces exactly what you’d expect: generic, interchangeable text that could apply to any marketing manager role at any company. The output is only as specific as the input.
Before you prompt anything, gather three things: the actual job description, one real achievement with a number attached, and one specific detail about the company. Feed all three into the prompt.
Prompts that actually produce something usable
I’m applying for [job title] at [company]. Here’s the job description: [paste it]. Here’s my most relevant achievement: [describe it with a specific number or outcome]. Here’s something specific about the company I want to reference: [detail]. Write a 3-paragraph cover letter — opening line referencing the company detail, one body paragraph built entirely around my achievement, and a closing that asks directly for an interview. Keep it under 300 words. Don’t use the phrases “passionate,” “dynamic,” or “team player.”
That last line matters more than it looks — naming specific words to avoid heads off some of the most obvious AI-cliché phrasing before it appears.
This is too generic in the opening and closing — rewrite both so they couldn’t apply to a different company. Keep the body paragraph mostly as-is.
Read this back as if you’re a hiring manager who reads 50 of these a week. Point out anything that sounds templated, overly formal, or like it was written by AI.
That third prompt is worth running even on a letter you wrote yourself — it’s a genuinely useful second-opinion check, not just a fix for AI-generated drafts.
The tells that give away an AI-written letter
A few patterns show up disproportionately often in unedited AI output — worth scanning for specifically before you send anything.
- Opening with “I am excited to apply for…” or “I am writing to express my interest in…” Both are technically accurate and say nothing — they’re the default the model reaches for absent a more specific instruction, and a hiring manager has read the exact phrase hundreds of times.
- The words “passionate,” “dynamic,” “thrilled,” or “leverage” appearing anywhere. These show up disproportionately in unedited model output because they sound confident without committing to anything specific — which is exactly why they read as empty to a human.
- A closing paragraph that restates the entire letter. Models tend to summarize before signing off, as if the reader needs a recap of three paragraphs they just read — cut it down to the direct ask instead.
- Claims with no specific example attached. “Proven track record of success” is a claim a model will generate confidently with nothing behind it, because it doesn’t have your actual achievement unless you fed it one.
- Perfectly symmetrical paragraph lengths. Real writing is rarely that evenly balanced — three paragraphs of nearly identical length is a subtle but consistent tell of unedited generated text.
None of these are disqualifying on their own, but two or more together is a strong signal the letter needs a real editing pass, not just a proofread.
Using AI to research the company, not just write
A separate, genuinely useful application: feed it a company’s “About” page or a recent press mention and ask it to summarize the two or three details most worth referencing in an opening line. This doesn’t replace ten minutes of your own reading, but it’s a fast way to process a long page you don’t have time to read closely.
Here’s the “About” section from [company]’s website: [paste it]. Summarize the two or three details from this that would be most specific and relevant to reference in a cover letter opening — avoid anything generic that could apply to most companies in this industry.
Verify anything it gives you before using it — a model can misread or slightly misstate a detail, and citing something inaccurate about the company in your opening line does more damage than a generic opener would.
The editing pass that actually fixes it
Three changes do most of the work:
Replace one generic sentence with one specific detail. If a sentence could apply to any company or any candidate, either cut it or replace it with something that couldn’t.
Read it out loud. AI-generated text often sounds fine silently and slightly off out loud — overly balanced clauses, a rhythm that doesn’t match how you actually talk.
Cut the last sentence of the closing paragraph. AI models tend to over-explain the close (“I am confident that my skills and experience make me an excellent fit, and I look forward to the opportunity to discuss further”). One direct sentence almost always beats two hedging ones.
What AI is genuinely useful for here — and what it isn’t
Good at: getting past a blank page, restructuring your own bullet points into paragraph form, catching grammar and awkward phrasing, generating a first pass fast when you’re applying to several roles in one sitting.
Not good at: knowing which of your achievements is actually the strongest one, writing something that sounds like your specific voice without heavy editing, or researching genuinely current, specific details about the company beyond what you feed it directly.
The output is only as good as the specificity you put in. AI can’t invent your best achievement or discover a genuine reason you want this specific job — that part is still yours to supply.
Once you have a draft you’re happy with, run it through the format and length checks in the complete cover letter guide, and confirm it still hits the standard margins, length, and structure rules — AI-generated drafts sometimes run long or drift from standard formatting conventions without anyone checking.
Or skip the back-and-forth entirely and build it directly in our cover letter builder, which starts you from a structured template instead of a blank prompt.
Ready to write your cover letter?
Write a job-ready cover letter in minutes with our AI-powered builder.
Create Cover Letter
FAQ
Is it okay to use ChatGPT to write my cover letter?
Yes, as a starting point — but an unedited AI draft is usually easy to spot, and it’s worth treating the output as a first draft that needs specific, personal editing, not a finished letter.
What should I include in an AI cover letter prompt?
The actual job description, one specific achievement with a number attached, and one detail about the company. A prompt with just a job title produces generic output because that’s all it has to work with.
How do I make an AI-written cover letter sound less like AI wrote it?
Replace generic sentences with specific details, read it out loud to catch unnatural rhythm, and cut over-explained closing sentences down to one direct ask.
Can a hiring manager tell if a cover letter was written by AI?
Often, yes — certain phrases and patterns (starting with “I am excited to apply,” words like “passionate” or “leverage,” an overly symmetrical structure) show up disproportionately in unedited AI output, and hiring managers who read many applications recognize the pattern.
Should I tell an employer I used AI to help write my cover letter?
There’s no expectation that you disclose this, the same way there’s no expectation you disclose using a spell-checker or a template. What matters is that the final content is accurate and genuinely represents you.
Does it matter which AI tool I use — ChatGPT, Claude, or Gemini?
Not much for this purpose. The prompts above work with any of them, and all three are prone to the same generic phrasing without specific input. The quality of what you feed in matters far more than which model processes it.
