Stop asking ChatGPT to rewrite your drafts. Try this prompt instead  | amznusa.com

So you’ve drafted a cover letter for a job application, but it just doesn’t feel right. Maybe your writing is a little flat (it’s hard to write about yourself, after all) or the letter just comes across as unfocused. You could go back for another editing pass, but ChatGPT is only a right-click away. 

Now, I’m not a fan of letting ChatGPT write for you. No matter how hard you try to steer ChatGPT, Claude, or Gemini toward your writing style, those notorious AI writing quirks (“it’s not this, it’s that!”) always manage to creep in. That’s why an AI-written cover letter will probably get your application dropped into the “no” pile.

That said, there’s a better way to get ChatGPT’s help when it comes to writing: making it act as a writing coach rather than simply ghostwriting for you.

Try this prompt on for size:

Here’s my [email / post / message / draft]. Don’t rewrite it. Read it as a thoughtful editor and tell me: what’s weak, what’s unclear, and where would a reader lose interest or get confused. Be direct — I want your honest assessment, not a polished version.

This prompt does a couple of key things. First, it stops the AI from simply optimizing your writing sample according to the statistical mean of what the most “useful” and polished version of it might be. AI chatbots like ChatGPT, Claude, and Gemini love to sand off the hard edges of any content, including writing, until they’re as smooth as possible, which is why AI writing often comes across as generic and “same-y,” all the way down to the em dashes and “it’s not this, it’s that” framing.

Secondly, the prompt makes the AI stop flattering you and zero in on what’s wrong with your draft, telling you bluntly where it’s vague, meandering, structurally unsound, or just plain boring.

I fed one of my recent articles to ChatGPT’s GPT-5.5 “Thinking” model using this prompt, and it gave me a blistering critique:

Your core point is useful and timely, but right now the article takes too long to get there, and the news setup overwhelms the practical lesson. Right now, the piece spends six paragraphs on “what happened / what might have happened” before saying “here’s the broader lesson.” That balance feels off. 

Ouch, but true. I left out the part where ChatGPT called one of my paragraphs “dense and bureaucratic” (ugh), while also catching a key logic issue: “If [Claude] Fable is overkill for most users, why should everyday users care that it disappeared?” Good point. 

This “critique, don’t rewrite” prompt is a popular one, namely because it’s so effective. Indeed, the writing critique you’ll get from this prompt is bordering on too harsh, which is why you might consider this prompt as a follow-up:

Now tell me which issue, if fixed, would do the most for the reader.

That add-on will help zero in on a critical writing issue, while also keeping you from feeling overwhelmed by the boatload of flaws ChatGPT just dumped on you.