How to use ChatGPT for your job search (without sounding like everyone else)

A recruiter friend in Munich showed me her inbox last Thursday. Forty-three applications for one mid-level marketing role. Twenty-nine of them opened with some version of: "I am writing to express my strong interest in the [Role] position at [Company], where I hope to contribute my passion for..." She didn't read past the second sentence on any of them. She's not being harsh — she's being honest. The default ChatGPT cover letter has a sound and a smell, and after a while you stop noticing the individuals behind them.

So the question isn't "should I use AI for my job search?" You should. Almost everyone is, and refusing to use the tools out of principle mostly just means you're slower. The question is how to use them so the output still reads like a person with a spine wrote it.

The failure mode: asking ChatGPT to write the thing

This is where 90% of people go wrong. They paste the job description in, type "write a cover letter for this," and ship the result. The output is generic because you asked for something generic. You didn't give the model any material to work with, so it filled the space with the average of a billion cover letters, which is exactly as bland as that sounds.

Everyone else is doing this. Stop.

What actually works: use it as a thinking partner, not a writer

Here's the flow that changed my own applications. I don't ask ChatGPT to write anything I'll send. I use it to do things I'm bad at or slow at.

Mining the job description. Paste the full listing and ask: "What are the three actual jobs-to-be-done behind this posting? What's the skill they're not quite saying out loud? What kind of candidate do you think they're tired of seeing?" You'll get ten ideas, three of which are useful. That's enough.

Stress-testing your story. Tell the model about your last role in messy, honest detail — what you built, what broke, what you were proud of, what you'd do differently. Then ask: "Which of these is the most specific and memorable? Which sounds like filler?" Models are good at this. They have no ego and will cheerfully tell you your second bullet point is boring.

Translating achievements into the company's language. If you're from e-commerce applying to fintech, the vocabulary is different. Ask the model to map your accomplishments into terms that would land with the fintech reader, then hand-edit. You're using it as a bilingual dictionary, not a ghostwriter.

A concrete example

A friend of mine — product marketer, 29, applying to climate tech companies after five years in SaaS — showed me three drafts of her cover letter. Draft one was pure ChatGPT: "I am excited to bring my passion for sustainability and my proven track record of driving growth..." I told her to burn it.

Draft two was hand-written, but dull. Real details, but arranged like a list.

For draft three we did this: she recorded a four-minute voice note talking about the launch she was proudest of at her last company. Messy, conversational, with the actual story of how the positioning nearly got killed in a product review and how she saved it. Transcribed it, fed it to ChatGPT with the prompt: "Here's a story I want to tell in a cover letter. Cut the filler but keep my voice. Don't smooth it out. Don't add corporate words."

The result was 70% there. She edited the last 30% herself — put back a specific detail the model had trimmed, cut a sentence that had crept into corporate-speak. The final letter opened: "The first time I argued with a VP about positioning, I thought I was going to get fired."

She got the interview. Two others that week too. The model didn't write her letter — it helped her find what was already in her voice and strip the fluff.

The tells to edit out

If you are going to use AI in your application, learn the giveaways and kill them on sight. The tells, as of this year:

• Any sentence starting with "In today's."
• "I'm excited to bring."
• "Proven track record."
• Triads of adjectives that all mean roughly the same thing ("strategic, driven, collaborative").
• Perfectly balanced sentence lengths throughout. Real writing has runt sentences. Three words. Like that.
• No specific numbers or names. If your cover letter could apply to four companies with a find-and-replace, it's too generic.

Read your final draft out loud before sending. If it sounds like a LinkedIn post from a stranger, it probably is.

Where AI genuinely helps more than it hurts

Fast interview prep. "Here's the job description and my CV. What five questions are they most likely to ask, and where am I weakest?" Gold. Do it the night before every interview.

Draft feedback. Write your cover letter first, without the model. Then paste it in and ask: "What's the weakest paragraph and why?" You don't have to accept the answer, but it's usually right.

Takeaway

Use the tools as a thinking partner and a ruthless editor. Don't use them as a ghostwriter. The goal is to apply faster and smarter without producing the same application as the 28 other people in the inbox. A little honest voice, one specific story, and ruthless trimming of model-speak will put you ahead of almost everyone.