How is AI already transforming jobs and skills?

Artificial intelligence is not making tech jobs disappear. It is gradually changing what they involve, automating certain tasks while increasing the value of other skills.

Take SAP administrators, for example. Just a few years ago, much of their time was spent monitoring systems, applying patches and performing system updates. Today, some of these tasks can be automated, but the role itself has not disappeared: it is evolving.

And this shift extends far beyond SAP. Across almost every tech profession, AI is changing the tasks that make up a role and, with them, the skills that make the difference.

Transformation, not disappearance

Recent studies paint a more nuanced picture than one of widespread job losses.

PwC’s 2026 Global AI Jobs Barometer primarily highlights the rapid evolution of the skills employers are looking for. The wage premium for AI skills has now reached 62%, up from 57% last year and as high as 118% in some sectors. PwC sees the labour market increasingly splitting into two paths: roles in which AI amplifies human expertise are growing faster and commanding higher salaries, while roles in which AI mainly simplifies tasks are stagnating. Similarly, according to a survey cited by EY, 74% of employees expect AI to change the nature of their work rather than eliminate their jobs.

The 2026 Cegos Barometer confirms the same trend: 68% of HR directors rank AI as the leading transformation ahead, while nearly one in four jobs could see some of its required skills become obsolete within the next three years.

That picture does, however, need to be qualified. Among junior developers, for example, the idea that AI is simply transforming jobs is harder to defend. According to Forbes, employment among junior developers has been declining for 33 consecutive months: the very tasks that traditionally served as an entry point into the profession are also among the easiest to automate.

Beyond whether certain jobs disappear or change, another question arises: how many people will be needed to perform the same amount of work? If AI enables teams to produce more, some companies may need to hire fewer people or operate with smaller teams. Job transformation can therefore also lead to lower overall employment levels, even when the jobs themselves do not disappear.

These developments show why it is more relevant to analyse the tasks that make up a job than to look at the job as a whole.

What we are seeing in practice

AI primarily automates repetitive or structured tasks. Conversely, interpretation, judgement and the ability to solve new problems are becoming more valuable. PwC finds that new tasks emerging in the occupations most exposed to AI are 2.5 times more likely to involve judgement, creativity or empathy.

This shift is already very tangible in the professions we work with every day.

In software development, adoption is already widespread: according to Google’s 2025 DORA Report, 90% of professionals use AI in their work. But adoption does not automatically translate into higher productivity. A controlled METR study found that experienced developers were actually 19% slower when using AI, even though they believed they were working faster. The DORA Report itself acknowledges that AI can amplify an organisation’s weaknesses just as much as its strengths. Trust also remains limited: 46% of developers surveyed by Stack Overflow say they do not trust AI-generated results.

Software development, however, is only one example. In data, queries and visualisations can now be generated using natural language. As a result, the analyst’s role is shifting towards data quality, interpretation and contextualisation.

The same logic applies to cloud environments. Tools such as SAP Joule can already automate certain maintenance tasks, while Basis professionals are increasingly moving towards supervision and architecture.

In cybersecurity, the transformation has an additional dimension: AI is both a defensive tool and a new weapon for attackers. It helps SOC teams automate alert triage, but according to IBM’s 2025 Cost of a Data Breach Report, one in six attacks already involves AI, primarily through phishing and deepfakes. Knowing how to use AI therefore also means knowing how to manage the risks that come with it.

All these examples point to the same conclusion.

Where the value is shifting

Across all these professions, value is increasingly concentrated in the ability to verify, interpret, make decisions and understand context. Cornerstone’s Global State of the Skills Economy report shows that demand for AI-related skills has increased by 245% in just one year.

But this growth is not limited to technical skills. Demand for emotional intelligence has risen by 95%, resilience by 42% and creative thinking by 18%.

At Triglav Digital, we see the same trend among our consultants: the profiles gaining the most value are those who know how to use AI without outsourcing their judgement to it.

If value is shifting towards new skills, professionals must also be given the opportunity to acquire them. This is probably where one of the biggest challenges for companies lies.

Adapting rather than reacting

Training remains the central issue. According to Cegos, only 32% of employees have received AI training, while 41% believe that the training they receive comes too late.

Since February 2025, this is no longer simply a matter of good practice. Article 4 of the EU AI Act requires organisations using AI to ensure that their employees have an adequate level of AI literacy, making it a genuine compliance issue for companies in Belgium and Luxembourg.

And that responsibility goes beyond training. It also includes shadow AI, tools used without formal approval or governance, as well as the confidentiality of the data processed through them, two areas that are still frequently overlooked in internal policies.

This raises an essential question: how should people be trained effectively? A meaningful learning path rarely consists of a single standalone e-learning module. It requires guided practice, dedicated time and sustained managerial support.

But the challenge does not stop with employees. As skills evolve, organisations also need to rethink job descriptions, performance criteria, the structure of their teams and, above all, the way they recruit. This last point is becoming particularly important at a time when AI-generated CVs and technical assessments are making it harder, rather than easier, to evaluate candidates’ actual capabilities.

For us, the real question is therefore no longer which jobs will survive AI, but which skills will continue to make the difference. For a consulting company like ours, this shift in value also raises questions about how we deliver our own services.

Let us return, finally, to our SAP administrator. The role has not disappeared. What has changed is how value is created within it. And that is precisely where the expertise of tomorrow will be defined.

Sources :