Polish employers increasingly expect workers to combine specialist knowledge with adaptability, independence and interpersonal skills. Data from the 20th edition of Gi Group Holding’s Labour Market Barometer show that soft skills are now the capability most often associated with higher productivity, ahead of specialist expertise and well ahead of technology alone.
The shift comes as artificial intelligence changes tasks across many occupations while Poland’s demographic outlook makes labour increasingly scarce. Employers are therefore looking not only for people who can use new tools, but for employees who can learn, make decisions and adapt when processes change.
Soft skills top employers’ productivity ranking
In the survey, 37.1% of companies identified soft skills as important for increasing productivity. Specialist knowledge followed at 34.9%, while 34.7% pointed to independence and responsibility for results. Flexibility and readiness for change were selected by 33.9%.
Skills employers associate with higher productivity
Share of surveyed companies
Soft skills — 37.1%
Specialist knowledge — 34.9%
Independence and responsibility for results — 34.7%
Flexibility and readiness for change — 33.9%
Source: Gi Group Holding, Labour Market Barometer, cited by Newseria Biznes.
The differences are small, which is itself important. Employers are not pointing to one dominant technical skill. Instead, they are describing a bundle of capabilities that allows employees to operate effectively in organisations where tools, workflows and business models are changing more quickly.
Nearly half of companies have struggled to recruit
Skills are becoming more important partly because recruitment remains difficult. Some 45.7% of surveyed companies said they had struggled in recent months to recruit at least one group of employees. The most frequently cited obstacle was candidates’ salary expectations, mentioned by 50.6% of companies facing recruitment problems.
The second-largest obstacle was a mismatch or insufficient level of skills, cited by 44.6%. This means the labour-market problem is not simply a shortage of applicants. In many cases, employers are unable to find people with the combination of knowledge, practical ability and behavioural skills they need at a salary both sides are willing to accept.
| Recruitment indicator | Share |
|---|---|
| Companies struggling to recruit at least one employee group | 45.7% |
| Excessive salary expectations* | 50.6% |
| Skills mismatch or insufficient skills* | 44.6% |
*Among reasons cited for recruitment difficulties. Source: Gi Group Holding data reported by Newseria.
Poland could have 2.1 million fewer workers by 2035
The demographic backdrop makes the skills issue more urgent. A simulation by the Polish Economic Institute cited in the report suggests that Poland’s labour market could shrink by almost 2.1 million workers by 2035. Fewer working-age people will increase pressure on companies to raise productivity rather than rely on continuous employment growth.
That changes the economics of training. When labour is abundant, companies can respond to changing needs by recruiting externally. In a tighter market, reskilling existing employees becomes more valuable because replacing them may be slower, more expensive or simply impossible in some occupations and locations.
AI matters, but employers do not see it as a substitute for broader skills
Artificial intelligence is part of this transition, but the survey indicates that companies do not see AI tools as a complete productivity strategy. Employees still need to understand when to use the technology, how to evaluate its output and how to connect it with a business process.
That distinction is becoming visible in corporate results. Research referenced by PARP in the Newseria report indicates that AI adoption is already widespread internationally, but only a much smaller group of companies can show a direct impact on profits. The gap between using an AI feature and redesigning work around AI remains substantial.
For Polish companies, this suggests that spending on licences or software without changing skills and processes may produce limited returns. The highest gains are more likely where employees understand the underlying business problem, can judge the quality of AI output and are able to reorganise their work around the new tool.
Employers expect partial automation more often than complete replacement
Another labour-market study cited in the report, the Human Capital Balance project, asked employers how automation could affect jobs over the next three years. Respondents expect 5% of current positions to be fully replaced by machines, robots or AI, while 22% could be partially replaced.
The distinction between full and partial replacement is crucial. In many occupations, AI is more likely to remove individual tasks than entire jobs. That increases the value of workers who can take over more complex responsibilities while delegating routine work to software.
The diploma is becoming a starting point rather than a guarantee
The broader message for employees is that formal education remains important but is less likely to guarantee a stable career path on its own. Knowledge acquired at the beginning of a career can become outdated more quickly, especially in technology-intensive occupations.
For employers, the same trend changes recruitment. Job descriptions built around a fixed list of software packages or narrow technical requirements may age quickly. Companies increasingly need people who can absorb new tools, communicate across teams and take responsibility for outcomes when the process itself changes.
In a country facing both demographic decline and rapid automation, those two pressures reinforce each other. Poland will have fewer workers, while each worker will operate with more technology. The competitive advantage will therefore come less from simply having access to AI and more from having people capable of using it productively.
Source: Newseria Biznes; Gi Group Holding Labour Market Barometer, Polish Economic Institute and Human Capital Balance data cited in the original report.
