Skills-based hiring: assessing a candidate against the job

What is skills-based hiring?

Skills-based hiring starts from one question: what will the person need to be able to do in this job, and how can we check it before hiring them? CVs and diplomas remain clues, but the decision rests on observable criteria, the same for all candidates.

This approach widens the talent pool to people whose background does not match the job title but who have the skills, for example career changers. It also makes the decision easier to justify, which matters under non-discrimination law.

How to assess a candidate against the job

1. Describe the job in terms of skills.
Start from the job's competency framework or, failing that, the corresponding sheet in France Travail's ROME 4.0 (the French public employment service's occupational directory). Select 5 to 8 decisive skills and distinguish those essential at hiring from those that can be acquired after arrival.

2. Define criteria and an assessment grid.
For each skill, write down what will be observed and at what level. The grid, set before applications arrive, is used by all assessors. It is also the best answer to the French Labour Code's requirement that recruitment methods be relevant.

3. Screen applications against these criteria.
Screening looks at elements linked to the job's skills. Criteria unrelated to them (age, address, photo, etc.) have no place in it.

4. Use work simulations.
An exercise close to real work (case study, simulation, work sample) tells you more than a question such as "Are you rigorous?". The exercise must remain proportionate and directly linked to the job.

5. Run a structured interview.
The same questions for all candidates, about past situations or practical cases, with each assessor scoring independently before pooling results.

6. Decide and prepare onboarding.
The gaps identified in the selected candidate are not lost: they become the starting point of their onboarding path and their first entry in the team's skills matrix, then in the skills map.

What is the recruitment by simulation method (MRS)?

The Méthode de recrutement par simulation (MRS) is a France Travail service that selects candidates on their "habiletés", meaning the abilities needed to do a job, rather than on experience or diplomas. According to France Travail, candidates take exercises designed to reproduce the job, in a session lasting half a day on average; those who succeed are presented to the employer without a CV, for an interview. The service is free, open to all recruiters, and covers about 240 occupations according to France Travail's presentation page. It is particularly suited to volume hiring and hard-to-fill jobs.

What does French law say about assessing candidates?

Text	What it provides
Labour Code, Art. L1132-1	No one may be excluded from a recruitment procedure on a discriminatory ground (origin, sex, age, family situation, health, disability, place of residence, etc.)
Labour Code, Art. L1221-6	Information requested from a candidate may only be used to assess their ability to do the job or their professional skills, and must have a direct and necessary link with the position
Labour Code, Art. L1221-8	The candidate is expressly informed, before they are used, of the recruitment methods and techniques applied; results are confidential and methods must be relevant to the purpose pursued
Labour Code, Art. L1221-9	No personal information about a candidate may be collected by a device that has not been brought to their attention beforehand
Labour Code, Art. L2312-38	The works council (CSE) is informed, before use, of recruitment methods or techniques and of any changes to them
GDPR, Art. 22	A person has the right not to be subject to a decision based solely on automated processing that produces legal effects or similarly significantly affects them, subject to regulated exceptions (including human intervention)

The CNIL, France's data protection authority, published a recruitment guide in 2023 with 19 fact sheets, including one on software for sorting, ranking and assessing applications, and another on potentially discriminatory data.

AI-assisted CV screening: what the EU AI Act changes

Regulation (EU) 2024/1689 on artificial intelligence (the AI Act) classifies as high-risk, in Annex III, point 4(a), AI systems intended for the recruitment or selection of people, in particular to place targeted job ads, analyse and filter applications and evaluate candidates. Point 4(b) also covers systems used for promotion decisions or to evaluate employees' performance.

A derogation exists for Annex III systems that do not pose a significant risk (for example a narrow procedural task), but a system that performs profiling of natural persons is always considered high-risk (Article 6(3)).

Timeline. The original text set 2 August 2026 for the application of the obligations for Annex III high-risk systems. Regulation (EU) 2026/1744, known as the "Digital Omnibus on AI", published in the Official Journal of the European Union on 24 July 2026, postponed the application of Sections 1 to 3 of Chapter III (requirements, obligations of providers and deployers) to 2 December 2027 for these systems.

Obligations of the employer as user ("deployer"). Article 26 requires in particular that human oversight be assigned to people with the necessary competence, training and authority; that input data be relevant and sufficiently representative where the deployer controls it; that automatically generated logs be kept for at least six months; and, before use in the workplace, that workers' representatives and the affected workers be informed. The regulation also provides, in Article 86, a right for the person concerned to obtain an explanation of the system's role in a decision affecting them.

The Défenseur des droits (France's rights ombudsman), together with the CNIL, warned as early as 2020 of the risk that algorithms reproduce discriminatory biases at scale, citing recruitment among the areas concerned.

Good practices for AI-assisted screening

• Keep a real human decision. The tool suggests, a trained person reviews and decides, with the ability to set the suggestion aside. A purely formal sign-off would leave the decision resting, in practice, on the solely automated processing covered by Article 22 of the GDPR.
• Configure it on the job's criteria. The criteria used by the tool must be those of the skills grid, explicit and directly linked to the job.
• Require explainability. For each ranked application, the assessor must be able to say which criteria weighed, so that a candidate can be answered and the decision justified.
• Check for bias. Regularly compare selection rates across relevant categories, in compliance with the GDPR, and report discrepancies to the provider.
• Inform. Candidates are informed that the tool is used (L1221-8 and L1221-9), and the CSE before it goes live (L2312-38).
• Document. Record of processing, impact assessment where required, provider's instructions for use, logs and reasoned decisions.