The majority of online applications today go through an algorithmic filter before reaching a recruiter. This observation changes the way a job search should be conducted: the choice of platform, the writing of the CV, and even the cover letter now respond to both technical and human logics. Understanding these mechanisms helps avoid the most common pitfalls and focus efforts where they yield results.
Automated application sorting: what the European regulation changes for candidates
The European regulation on artificial intelligence classifies tools used to filter applications or evaluate candidates as high-risk systems. This classification imposes transparency and traceability obligations on employers regarding the algorithms they use.
For a candidate, the direct consequence is twofold. First, check if the job offer mentions the use of an automated sorting tool. Second, adapt their application so that it is readable by both a machine and a human.
French law goes further: an automatically rejected application can lead to a request for human re-examination. The GDPR regulates exclusively automated decisions that produce a significant effect on a person. In practice, this means that a candidate rejected by an algorithm can request an explanation, contest the decision, and have a human being reassess their file.
Few candidates exercise this right, often due to lack of knowledge. Consulting employment services on Services Emplois allows access to offers where the application process remains identifiable, making this type of approach easier if a rejection seems unjustified.
Online job search: adapting your CV to algorithmic filters without losing readability
Applicant Tracking Systems (ATS) analyze CVs based on keywords extracted from the job description. A visually elegant CV structured in complex columns or embedded images risks being poorly read by these tools.
Three technical principles improve a CV’s compatibility with these filters without sacrificing quality for a human reader:
- Use the exact terms from the job offer in the skills section, rather than synonyms or creative rephrasings. An ATS looks for literal matches.
- Use a simple format (a single column, explicit section titles like “Work Experience” or “Education”) and favor text PDF format, not image PDF.
- Place the most relevant information in the first third of the document. Some ATS stop analyzing after a certain volume of text.
An ATS-optimized CV remains a clear CV for a human recruiter. The two objectives do not oppose each other, contrary to what some guides suggest that push to stuff the document with invisible keywords. This technique, sometimes called “keyword stuffing in white on white,” is detected by recent tools and can lead to automatic rejection.
AI-generated cover letters: the concrete risks of an unreviewed application
Generative AI tools produce cover letters in seconds. The time-saving is real, but the generated content may include discriminatory or inappropriate formulations without the candidate realizing it.
From the employer’s side, the responsibility remains complete: a job offer written by an AI assistant that implicitly excludes certain profiles exposes the company to litigation. For candidates, the risk is different but equally concrete. A generic letter, recognizable by its repetitive structure and catch-all phrases, reduces the chances of being read beyond the first paragraph.
Field feedback varies on this point: some recruiters claim to immediately spot a generated text, while others admit they cannot distinguish if the content is relevant. However, there is consensus that personalization remains the decisive criterion. Mentioning a specific project of the company, an identified sector issue in the offer, or a skill directly related to the position transforms a generic letter into a memorable application.
Job offer platforms: selection criteria beyond notoriety
Multiplying registrations on all available platforms gives the illusion of an active search. In reality, focusing efforts on two or three platforms suited to one’s sector yields better results.
Criteria that deserve verification before registering:
- The freshness of published offers. Some platforms display expired or already filled ads, which skews the perception of the market.
- Transparency regarding the recruitment process. A platform that specifies whether the employer uses an automated sorting tool allows the candidate to calibrate their application.
- The ability to set alerts by skills rather than by job title. Job titles vary from one company to another, but the sought skills remain stable.
- The quality of information about the company (size, sector, recruitment policy). A detailed employer profile reduces the time spent on additional research.
Aggregators compile offers from multiple sources, which broadens the spectrum. However, the duplication of ads can skew the tracking of applications if the candidate does not centralize their efforts in a personal management tool, even a simple spreadsheet.
Application tracking: structuring your search to avoid dispersion
Sending applications without keeping track generates inappropriate follow-ups or forgetfulness. A structured follow-up does not require a complex tool: a table with the date of sending, the position, the company, the channel used, and the status of the application is sufficient.
Following up on an application between seven and ten days after sending remains a practice recommended by most recruitment professionals. Beyond this timeframe, the offer is often in advanced selection stages.
Tracking also helps identify patterns: if a majority of applications sent via a particular channel go unanswered, the problem may lie with the channel itself (outdated offers, overly restrictive filters) rather than the candidate’s profile. Adjusting one’s job search strategy based on this data, even rudimentary, avoids repeating the same mistakes for weeks.
