ChatGPT for Swiss Job Applications: Why Generic AI Letters Get Flagged — and How to Do It Right

Asking ChatGPT to "write a cover letter for this job" has become one of the most common shortcuts in job hunting — and recruiters know it. That alone is not a problem: using AI to draft application documents is neither forbidden nor shameful, and plenty of strong candidates do it. The problem is generic AI output, and in Switzerland it carries fingerprints that a hiring manager spots within seconds.

The most visible one for German-language applications: large language models are trained overwhelmingly on German-from-Germany text. Left to their defaults, they write ß where Switzerland writes ss, close with "Mit freundlichen Grüßen" instead of "Freundliche Grüsse", and quote salaries in € instead of CHF. Add interchangeable stock phrasing and the occasional invented qualification, and the letter that was supposed to save you an evening quietly costs you the interview.

Why generic AI letters get noticed

Recruiters read stacks of cover letters, and since 2023 an ever-growing share of them come from the same handful of AI models prompted the same lazy way. The result is convergence: dozens of letters opening with "I am writing to express my strong interest in..." or, in German, "Mit grossem Interesse habe ich Ihre Stellenanzeige gelesen", followed by three paragraphs of adjectives ("dynamic", "results-oriented", "passionate") that could describe any candidate for any job at any company.

To be fair: none of these phrases is wrong. Human applicants wrote them for decades before AI existed. What changed is the density. When a model generates text without specific input, it falls back to the statistically most likely phrasing — which is by definition the phrasing everyone else gets too. A cover letter has exactly one job, and that is to differentiate you. A letter assembled from the most probable sentences does the opposite.

The tell is rarely one sentence. It's the pattern: perfect grammar, confident tone, and nothing that could only have been written about you and this specific company. Swap the company name and the letter still works — that's the test recruiters intuitively apply, and generic AI output fails it every time.

We covered the human version of this problem in our guide to cover letters that don't sound generic; AI simply industrializes it.

The Germany-German default

For applications in German-speaking Switzerland there is a second, more mechanical layer of tells. AI models default to the conventions of Germany, not Switzerland — and Swiss readers notice immediately, because these are exactly the details Swiss schools and employers drill.

None of these mistakes says "bad candidate". But together they say something arguably worse in the Swiss context: this person did not look closely. An application containing ß and € signals a copy-paste job — or a candidate blasting the same letter across Germany, Austria and Switzerland. For a market that prizes precision and preparation, that first impression is hard to recover from.

The same logic applies in reverse for French and Italian: models default to France-French and Italy-Italian conventions and euro amounts, which reads as equally unlocalized to a recruiter in Lausanne or Lugano.

Hallucinated qualifications and empty mirroring

The subtler failure mode has nothing to do with spelling. When ChatGPT writes a letter without your real CV in front of it, it fills the gaps with plausible-sounding material: "five years of experience in project management", proficiency in tools you've never opened, even certifications you don't hold. The model isn't lying — it's completing a pattern. But you are the one signing the letter.

This fails in two places. First, recruiters read the cover letter next to the CV, and mismatches jump out: a letter praising your "extensive leadership experience" attached to a CV showing none is worse than no letter at all. Second, everything in the letter is fair game in the interview. A claim you can't back up with a concrete example collapses in the first ten minutes.

The mirror-image problem is a letter with no grounding in the actual job ad: the model reflects the job title back and pads around it. Swiss hiring culture expects the opposite — a letter that shows you read the posting, understood what the role actually involves, and can name a specific reason why this company.

What Swiss recruiters expect instead

The Swiss application dossier follows well-established conventions, and the cover letter — usually called a Motivationsschreiben — is only one part of it:

• One page, no more. Concise beats comprehensive. Three to four focused paragraphs that connect your experience to this role.
• Swiss orthography and currency. ss instead of ß, "Freundliche Grüsse", amounts in CHF.
• A complete dossier. Letter + CV (1–2 pages) + Arbeitszeugnisse (employer references) + diplomas, ideally combined into a single PDF. Swiss employers genuinely read the Arbeitszeugnisse — a polished letter cannot compensate for a missing dossier. Our Swiss CV guide covers the CV half in detail.
• Specific references to the company and role. Name the position, refer to something concrete from the posting, and give one real reason you fit that a competitor's applicant could not copy.
• Your actual profile, accurately. Nationality/permit status on the CV, claims in the letter that match the CV line for line.

Notice what's not on the list: poetic language, clever hooks, or maximum keyword density. Swiss recruiters reward precision and relevance, not rhetoric — which is good news, because precision and relevance are exactly what you can feed an AI.

How to use AI well

Used properly, AI is a legitimate and effective drafting tool. The difference between a flagged letter and a strong one comes down to what you put in and what you check on the way out:

1. Feed it your real CV. Paste your actual work history, or use a tool that has your structured CV data. The model can only be accurate about you if it knows you.
2. Feed it the real job ad — the full text, not just the title. The letter should respond to the stated requirements, not to a stereotype of the role.
3. Ask for Swiss conventions explicitly — Swiss spelling (ss, never ß), "Freundliche Grüsse", CHF — or use a tool that applies them automatically.
4. Verify every factual claim. Years of experience, tools, certifications, language levels: check each one against your CV before sending. Delete anything you couldn't defend in an interview.
5. Add the one thing AI can't know: your genuine, specific reason for applying to this company. One honest sentence beats three generated paragraphs of enthusiasm.
6. Rewrite the opening and closing in your own voice. These are the most template-prone parts and the ones recruiters read first and last.

If a draft passes the swap test — replace the company name and the letter stops making sense — you're in good shape.

Where preparAItor fits

preparAItor exists because doing all of the above manually, for every single application, is exactly the evening of work AI was supposed to save. So the platform bakes the checklist in:

• Grounded generation, by design. Documents are generated from your parsed CV plus the actual job posting you provide (URL, PDF, or text) — not from a generic prompt. The model literally cannot invent a qualification that isn't in your CV data, and the letter responds to the real requirements of the real ad.
• Swiss orthography toggle. For German output, one setting writes ss instead of ß throughout — no find-and-replace, no "Grüße" slipping through.
• Swiss-market tones. A Formal tone suited to Swiss conventions is included on every plan; Pro and Premium add four more to match different company cultures.
• DE, FR and IT on all plans — including Free — so you can apply across language regions without importing another country's conventions. English output is available on Pro and Premium.

The free plan includes 250 credits per month with no credit card — roughly two complete application packages — so you can compare the output against a generic prompt on a real job ad before deciding anything.

The honest conclusion: ChatGPT is not the enemy, and neither is AI in applications. The enemy is the generic, unlocalized, ungrounded letter — whoever or whatever wrote it. Ground the text in your real CV and the real posting, localize it for Switzerland, and keep your voice in it, and AI becomes what it should have been all along: a way to apply better, not just faster.

Sources

• Swiss Federal Chancellery — Schreibweisungen (official writing rules for federal German texts)