Blog 22 March 2026 · 4 min
An AI's admission: why generic tools sabotage your job application
Are you using ChatGPT to write your cover letter to save time? Plenty of candidates do, but it is also a trap that sabotages your chances. A generic AI is built to rephrase, not to analyse in depth how well your CV fits a job advert. The result? Bland, impersonal text that recruiters and ATS spot and discard straight away. According to stubby-ai.com, applying for a job is not a writing exercise but a demonstration of strategic analysis. This article explains why general-purpose AIs fail and how a specialised AI becomes your best asset for building a compelling application and landing interviews.
"I can't do this task": the admission that gives it all away
Picture the scene: you open ChatGPT, upload your CV, paste in a job advert and ask: "Write me a powerful cover letter." The answer (sometimes explicit, often implicit):
"I cannot carry out this task as it is phrased. I am designed to analyse and summarise only the search results I have been given, not to carry out new independent research."
Not a bug. Not sabotage. A blunt truth that is rarely spoken: general-purpose AIs are not built to investigate, cross-reference and build a tailored strategy. And that is where everything goes wrong (for you).
What recruiters expect: analysis, not copy and paste
People think applying for a job means writing. It doesn't. It means analysing.
✔️ You need to understand and break down the advert: keywords, duties, hidden challenges, company values.
✔️ Compare your CV with those expectations: where have I already proved what they are asking for? What am I missing?
✔️ Bring the two together: build a story, an argument. Why am I THE solution to their problem?
This work is what we call a "business case" application: you offer a genuine, made-to-measure solution. Exactly what the ATS (automated screening software) or the recruiter wants to see: an accurate answer, not one-size-fits-all text.
Why ChatGPT and generic AIs fail (and sink your application)
I'll be blunt: asking ChatGPT to "write me a cover letter for this job with this CV" is like thinking you're making chocolate mousse when half the ingredients are missing.
ChatGPT is designed to rephrase, summarise and make things look nice. But it does not cross-reference two texts to analyse the common ground, the gaps and the finer matches. It takes bits from one side, adds sentences from the other, sprinkles in some pleasantries: it sounds good, but it does not respond strategically to the advert.
The result? Letters that all look the same, that dodge awkward topics, that never put YOUR background in the spotlight for THIS specific role. In short: generic text that reeks of AI and burns your chances with recruiters, as shown here.
This is exactly what the Harvard Business Review points out: generative AIs are good at "dressing up content" but fail to reason and cross-reference multiple sources to deliver personalised analysis; they produce the plausible, not the relevant source: HBR.
Generic AI vs specialised AI: the difference that changes EVERYTHING
Imagine a universal coffee machine. It can make coffee, tea, hot chocolate… But ask for a really strong espresso at exactly the temperature you like? Too specific, too personal. That is ChatGPT in the job search: strong on "form", weak on substance, nuance and cross-referencing.
By contrast, a specialised AI like Stubby is an Italian espresso made to order.
Why?
- It breaks the advert down into duties, skills and criteria.
- Your CV is scanned line by line, with strengths and gaps analysed.
- The cross-referencing is methodical and automated: areas of fit and mismatch are laid bare.
- The result is a report and texts that are 100% tailored. No waffle, just solid arguments. So you can compete on a level playing field.
That is the big lesson highlighted by every study on the subject, like this one: there is a fundamental difference between a "general-purpose" AI that makes things look nice and a "specialist" AI built for analysis, matching and genuine personalisation.
This recent comparison reaches the same conclusion: ChatGPT (or Gemini, Copilot, etc.) can generate text quickly, but for extraction and compatibility analysis you need a specialised AI (read here).
The real solution: equip yourself to meet the recruiter's analysis
Recruiters use tools that decode your application and analyse how closely it matches and what concrete evidence it offers. This is not a myth.
Faced with that, competition is fierce:
Either you send plausible but impersonal text. Or you equip your application with an AI that does this in-depth analysis for you.
Stubby does not just produce a letter: it generates an objective report that explains the gaps, highlights your strengths and even gives you the arguments to put forward to convince this recruiter, at this moment!
That is the difference between keeping up appearances… and winning the battle.
So remember: applying for a job is not a test of style but of analytical intelligence. And for that mission, ChatGPT is not enough.
So, ready to move to a truly strategic application?
Try it for free at stubby-ai.com and give yourself every chance, without wasting time.
Main keywords: generic AI, specialised AI, application analysis, CV optimisation, Stubby AI
« When facing a recruiter who analyses, don't turn up armed with just a piece of text, but with a strategy. »
| Section | Key messages |
|---|---|
| The AI's admission | ChatGPT acknowledges its limits: unable to cross-reference and analyse CV + job advert the way a human can |
| The candidate's real job | Applying is not just writing; you need to analyse the advert and your background, then link them precisely |
| Why generic AI fails | Mainstream AIs produce "plausible" but superficial text, offering no evidence of fit |
| Strengths of specialised AI | Only an AI designed for matching (like Stubby) can cross-reference, analyse, recommend and personalise at this level |
| Result | A credible, competitive application that anticipates the recruiter's analysis and maximises your chances of an interview |
