Tools

AI-generated text

Skills-based assessment could improve recruitment matches

Hungary’s labour market appears broadly stable, but artificial intelligence and a weaker economic environment are changing recruitment, according to labour market specialist Horváth Tamás.

Skills-based assessment could improve recruitment matches

Hungary’s labour market shows no dramatic decline in the statistics, but substantial structural changes are taking place beneath the surface, according to Horváth Tamás, a labour market and organisational development specialist and managing director of HUN-TER-APP Kft. Data from the Hungarian Central Statistical Office (Központi Statisztikai Hivatal, KSH) show that fewer people are working than two years ago and that the share of long-term jobseekers has risen, while demand and uncertainty vary across occupations.

Horváth says the effects of artificial intelligence are already being felt in white-collar, knowledge-based roles. Europe’s economic slowdown, difficulties in the automotive industry and a more uncertain business environment are also affecting the Hungarian market. Companies are hiring more cautiously, recruitment can take longer, and many are filling only positions they are certain they need.

He argues that skills, learning ability and adaptability are becoming increasingly important alongside qualifications and previous job titles. Yet a conventional CV often reveals little about how a candidate thinks, how quickly they learn or the environment in which they perform best.

Clearer requirements and more visible skills

According to Horváth, recruitment problems often begin with the definition of the vacancy. Employers do not always distinguish between knowledge needed from day one and abilities that can be developed on the job. Requirements that are too broad can attract many unsuitable applicants, while overly narrow conditions can exclude capable people from less conventional backgrounds.

He says employers should clarify working conditions as well as tasks: the degree of autonomy required, how the team operates, the management support available and the collaboration expected. Fit is therefore not simply a matter of personal rapport, but of assessing the conditions under which a candidate can work effectively.

HUN-TER-APP Kft. is developing the HUN-TER platform to support this process. Horváth says the planned system would help companies define tasks and competencies more precisely, while allowing candidates to present their experience, knowledge and development potential in structured professional profiles. Its aim is not to let AI make hiring decisions, but to reveal connections, missing information and questions requiring clarification.

For example, the system could help identify experience relevant to a vacancy even when candidates and employers use different terminology. Horváth stresses that AI cannot replace verification. Because generative AI can produce a convincing CV, it is particularly important to distinguish self-reported claims from experience supported by projects, qualifications or other evidence. Detailed profiles could also help identify possible training and career paths.

Limits of AI recommendations

EU regulation treats AI systems used in recruitment as a high-risk area. Horváth says a percentage match score can be misleading on its own: a result of 74 or 82 percent does not mean a candidate has the same probability of succeeding in the job. The figure depends on the criteria, weightings and data used.

He says systems should also show which matches are supported by information, where differences exist and where evidence is insufficient. Missing data must not automatically be treated as a missing skill. Horváth argues for transparent systems that explain their reasoning and uncertainties, with the final decision made by a person.

Overlooked labour market potential

Companies vary in their readiness to assess candidates this way. Horváth says larger businesses with deliberate HR strategies are more likely to examine competencies, motivation and learning ability, while smaller firms often lack the capacity. A conventional interview can also mislead: a candidate who freezes when asked a factual question might perform well based on other skills if the role does not require a rapid response under stress.

He sees overlooked potential among people in their forties and fifties, among others. Experienced workers may struggle to describe customer service, problem-solving or relationship-management skills in the language of contemporary job advertisements, and may not recognise which roles suit them. The strengths of early-career applicants, career changers and people returning after a long absence can also remain hidden in a conventional CV.

Horváth says more detailed profiles can make skills, education, practical knowledge and development potential visible alongside formal work experience. This could help identify suitable jobs and career directions.

Changing HR responsibilities

AI can support repetitive administrative and initial screening tasks, such as reviewing large numbers of CVs and comparing information. Horváth nevertheless believes that personal conversations, building trust, understanding motivation and assessing organisational fit will remain human responsibilities.

He says the technology’s effectiveness should not be measured solely by faster administration. Employers should also examine whether it identifies more relevant candidates, improves decision quality and affects employees’ subsequent performance, satisfaction and retention. Pilot projects, real-world experience and consistent measurement can provide answers.

Horváth Tamás is scheduled to speak at the HR (R)EVOLUTION conference on November 17, 2026. Publication of the original article was supported by Hun-Ter App Kft.