AI for your job application

Most people throw their old CV into ChatGPT and write "make this better". The result then reads like a LinkedIn post from someone who doesn't exist. Full of words like "results-oriented", "dynamic" and "passionate". No recruiter believes it, and you notice yourself that it doesn't sound like you.

Yet a job application is exactly the kind of task where an AI tool can help really well. You just have to know what to use it for and what not. In this lesson I'll show you how I'd split it up: what the AI does, what you do, and where the line is.

Why AI actually helps with applications

Applying is a slog, because you have to pour the same information into five different forms. A CV, a cover letter per position, a LinkedIn profile, maybe a short intro-video script. And every time you wonder whether you're getting your strengths across right.

This is exactly where AI is strong. Not because it's more creative than you, but because it can quickly rephrase, shorten, structure and play through perspectives. It knows thousands of job ads and knows which phrasings recruiters expect. What it can't do: know your real story. That's on you to provide.

Think of the relationship like an editor. A good editor doesn't write your book. They polish, cut, ask questions. That's exactly how you should use the AI.

Step one, gather the raw material

Before you open any tool, you need material. Sit down and write it down, however messy:

Your recent roles with years. What you concretely did there, not the job title but the actual work. One number per role if possible, so "team of four", "budget of 20,000 euros", "cut the complaint rate from eight to three percent". Numbers like that are gold, because the AI isn't allowed to invent them and a recruiter loves them.

That sounds like work, but it's the most important step. The quality of your application depends almost entirely on what you put in. Feed the AI nonsense and you get polished nonsense out. In the last lesson you learned that a model only works as well as its context. That applies exactly here.

Step two, sharpen the CV

Now the tool comes into play. Take ChatGPT or Claude, it doesn't matter for this task. Put in your collected raw material and phrase your request concretely. Not "make a CV", but something like:

"Here are my career stops as bullet points. Turn them into punchy CV entries. Maximum three lines per role, active verbs, no clichés like results-oriented or team player. Keep all numbers exactly as they are."

The last sentence matters. AI tools tend to round numbers or embellish them if you don't rein them in. If you write "eight to three percent" and suddenly it says "significantly reduced", then you've lost your best piece of evidence.

Read the result critically. Cut everything you couldn't back up with a story in the interview. If it says "strategic realignment of sales" and you were actually only in two meetings, then that sentence goes. Honesty isn't a moral extra, it's self-protection. Every inflated point is a question you can't answer in the interview.

Step three, the cover letter per position

Here most people make the mistake of building one cover letter and sending it fifty times. Recruiters smell that immediately. The cover letter is the place where you show that you've understood this one specific position.

Paste the job ad into your tool along with the CV you just sharpened. Then ask for something like:

"This is the job ad, this is my background. Find the three requirements from the ad that best match my experience. Suggest a concrete sentence for each that connects my experience to the requirement."

Notice the difference? You're not having the whole cover letter written, but the bridges between your profile and the position. You then assemble those bridges into a text yourself. That way your voice stays in it, and the text still fits the position.

A cover letter that comes entirely from the AI sounds smooth and empty. One where the AI only gives you the connection points and you write the rest sounds like you. That's the point where many people give up and let the machine do it. Don't.

Step four, rehearse the interview prep

This is the underrated superpower. You can use the AI as a practice partner for the interview. Give it the job ad and your CV and ask it to ask the ten most likely interview questions. One after another, not all at once.

Then you answer. Honestly, the way you would in the real interview. After each answer you can ask "where was my answer weak, what would a recruiter dig into". The AI is a patient sparring partner that's never annoyed and doesn't laugh at you.

Especially useful with the uncomfortable questions. The gap in the CV, the job change after eight months, the salary expectation. Play these situations through two or three times until you have an answer that feels honest and calm. Not memorized, but thought through.

The line where AI becomes harmful in an application

There are moments where you should close the tool. First, with anything that's personal motivation. When the cover letter says why exactly this company appeals to you, that has to come from you. An AI-generated enthusiasm for the "innovative company culture" is transparent and embarrassing.

Second, with sensitive data. If your application contains health information, family details or the exact salary from your current employer, think hard about what you type into a tool. Data protection when using AI is a topic of its own, and when applying you're right in the middle of it.

Third, with the final tone. Read every application out loud once before you send it. If a sentence feels like it could come from anyone, then cut it or rewrite it yourself. Your goal is not the perfect application, but one that sounds like you and is still professional.

What you've got now

You now know that when applying, the AI is a tool and not a ghostwriter. It gathers, structures, rephrases and plays through scenarios. The substance, the honesty and the voice come from you. It's exactly this division that separates an application that stands out from one that disappears in the pile.

If you want to take the next step, check out the lesson on hallucinations in Level 1. Because precisely with numbers and facts in your application, it's crucial that you recognize when the AI invents something. And you should look at the topic of data protection when using AI before you dump personal documents into a tool.