Your CV can be perfect. But did you actually look at us?

A flawless CV and a convincing cover letter used to be important signals in a job application. Today, they may mainly show that someone knows how to work with a text generator/LLM.

That does not mean we should distrust candidates because they use AI. A candidate may use AI to improve their language, structure their thoughts or lower a barrier that has nothing to do with their ability to do the job. But when every application sounds professional, enthusiastic and error-free, it becomes harder to see who has genuinely paid attention to the organisation and the role.

A recent Tilburg University study, published in the Journal of Labor Economics in 2026, compared job applications written with and without ChatGPT. AI mainly improved generic elements such as spelling, formal structure and opening and closing sentences. According to the researchers, that did not increase the chance of receiving an interview invitation. Recruiters cared most about personal motivation and clarity, and those elements did not automatically become stronger through AI.

The problem is not AI. It is impersonal output.

A candidate who uses AI to make a text easier to read is not necessarily a less motivated candidate. The signal we are looking for as employers lies somewhere else: did someone try to understand who we are, which problem we are solving and why this particular role is interesting to them?

That distinction matters more as generic application texts can be produced at scale. A cover letter can mirror the words from a job description perfectly and still tell us very little about the person sending it.

That is why we would rather ask candidates a few concrete questions than ask them to write even more. For example:

• What attracts you to this role? Not the job title, but a specific part of the work.
• Which problem of ours do you think you could help solve? The answer shows whether someone looked beyond the job description.
• What would you want to understand first if you started tomorrow? This creates room for curiosity and meaningful preparation.

A short, role-specific quiz is not an automatic selection machine, nor is it a test of general intelligence. It is a low-friction way to see whether someone has taken the time to understand the role and the organisation. It also gives candidates a fairer opportunity to show their interest than a generic cover letter that is mainly judged on style.

First commitment, then a deeper look.

A short quiz can improve the first signal, but it cannot tell us how someone will act in a real work situation. That is why we should keep this step proportionate. The quiz can check whether someone understands the context and is making a deliberate choice to apply. After that, a recruiter can invite the most relevant candidates to a conversation or a deeper assessment stage.

The questions themselves also benefit from structure. In 2019, Hartwell, Johnson and Posthuma examined four types of structured interview questions with 303 applicants. In their sample, background questions, situational questions and questions about past behaviour predicted later job performance. The lesson is not to create longer questionnaires. It is to ask concrete, job-relevant questions consistently.

From polished words to observable behaviour.

Once candidates have cleared that first step, we want to understand more than why they found the vacancy interesting. We want to understand how they make choices, approach problems and respond when new information changes their first approach.

At Assess4me, we use game-based assessments for that purpose. Not to give candidates a mysterious score or to take the decision away from the recruiter, but to make behaviour observable in an additional way. Recruiters can use the results as a starting point for better questions and a richer conversation with the candidate.

The future of selection does not have to be made up of longer forms or increasingly clever text generators. It may start more simply: with a short question that only someone who has paid real attention can answer well, followed by an assessment that shows how that person thinks and acts.

AI can make applications look better. That does not automatically make them more informative. In a labour market where words are becoming cheap, genuine preparation becomes valuable again.