Artificial intelligence will change not only the way current employees work, but also the way younger people enter the labour market and take their first steps in their careers, particularly in the tertiary sector and the service industries. The challenge will be much greater for young people because the tasks through which one learns a trade generally involve researching information, preparing documents, carrying out preliminary analyses, basic programming and support activities. With AI, analyses and documents become more abundant and easier to find, although verifying their reliability then becomes essential. The value of skills that are more difficult to automate is on the rise, such as building trust, understanding the context and taking responsibility for decisions made with the support of AI. And within this challenge lies another, explains Nicola Mauri, head of Valeur Group for Italy (an independent financial group active in wealth management, asset management, the structuring of investment vehicles, trading and property), the challenge within the challenge is ‘building a relationship of trust between AI and the organisations that hold expertise, processes and relationships’. For those joining a company today with an almost ‘innate’ familiarity with AI, ‘acting as a bridge between the new possibilities offered by AI and the roles that require greater experience and relationships could become a new first step in their career. The question, therefore, is not merely how many jobs AI will replace, but where it can actually be integrated into processes and how the resulting productivity gains will be utilised: to reduce the workload or to create new tasks, skills and career paths?”
The paradox is that AI can eliminate some of the early stages of a career and, at the same time, create new ones, helping new entrants to progress through them more quickly. Data from a study carried out by the Valeur Group show that over half (51 per cent) of the organisations surveyed report a reduced need for entry-level roles and, ‘among workers aged between 22 and 25 in occupations most exposed to AI, a relative decline in employment of around 13 per cent has been observed. At the same time, in customer support, generative AI has increased productivity by 14–15 per cent, particularly amongst less experienced workers’, continues Mauri.
Loading…
This was the starting point for the “Beyond Exposure” analysis carried out by the company to understand what is happening and to create a model for interpreting the current context. “Knowing the extent to which a job is exposed to AI is not enough to predict its impact on employment. What matters is the extent to which individual tasks are actually ‘convertible’ into production processes – that is, digital, repeatable, verifiable and characterised by an acceptable cost of error,’ argues Mauri. The research findings already reveal significant differences between sectors and countries. In the United States, redundancies in the IT sector have risen by 26 per cent and the employment trend has been negative, with around 7,700 jobs lost per month. By contrast, education and healthcare have continued to create over 51,000 jobs per month. “In our model, it is precisely IT and sectors linked to data analysis that are the sectors in which exposure and convertibility are highest,” says Mauri.
Italia, too, provides concrete evidence of this trend, according to the Valeur Group. According to Istat data from 2025, amongst Italian companies using AI, seven out of ten employ it to extract information from text documents, whilst only 5.9 per cent use it in applications related to the physical movement of machines; workflow automation stands at around 18 per cent. ‘AI is therefore making its way first into areas where work is already digital and verifiable, in line with the concept of “convertibility” at the heart of our research,’ explains Mauri, who adds, however, that a generalised reduction in employment cannot yet be observed. “In a recent global survey conducted by the company, 32 per cent anticipated AI-related job cuts a year ago. Twelve months on, only 14 per cent report that these have actually taken place, whilst around two-thirds report no changes or only limited changes. AI therefore seems to be changing the way we work and the mix of skills more rapidly than the overall number of jobs.”
According to Mauri, to understand where change will occur first, one must assess the degree to which a job can be automated. “In our model, the feasibility rate stands at 82 per cent for a software developer and 85 per cent for a data analyst. It drops to 52 per cent for a nurse and to 30 per cent for a C-suite executive,” says Mauri. “The difference is not just technological: where value depends on judgement, responsibility, context and personal relationships, automation is much more difficult. This is where the role of trust and human oversight becomes crucial. In our model, if an error is difficult to correct or has significant consequences, entrusting the process entirely to a machine becomes less acceptable. This is one of the reasons why healthcare, teaching, leadership and relationship management prove more difficult to replace with AI.”