The Swiss IT job market has been presenting us with a curious paradox. Unemployment among IT professionals is rising, yet employers continue to report hiring difficulties, while forecasts still warn of a shortage of skilled workers.
There is no shortage of explanations: unrealistic requirements, salary caps, international competition, mismatches between education and market needs, age discrimination, outsourcing, CV screening, or simply a gap between the skills available and those employers are looking for.
All of these probably play a part. But they may also be obscuring a deeper shift: companies still need the skills. They may simply no longer need the people who have them on a full-time basis.
From in-house expertise to expertise on demand
IT has been moving in this direction for years.
A company that once ran its own servers, networks, applications and databases naturally employed the people required to operate them. Cloud computing, SaaS, managed services and outsourcing gradually changed that model.
The expertise did not disappear. It moved.
The DBA now works for a vendor. Network expertise sits with the MSP. An architect joins a project for a few days. The Microsoft 365 specialist is brought in when needed. The migration lead arrives for several months and moves on to another client once the job is done.
Taken individually, these decisions make perfect sense. Why employ a specialist full time if you only need forty days of their expertise each year?
The problem emerges when the same reasoning is repeated year after year.
Eventually, the organisation discovers that it no longer has certain capabilities in-house. When a complex project comes along, nobody internally has the necessary experience. So an external expert is brought in.
And because complex projects keep going to external experts, internal employees get fewer opportunities to acquire that experience themselves.
The cycle becomes self-reinforcing:
less expertise in-house → more outsourcing → fewer opportunities to build expertise internally → greater dependence on external specialists.
It is therefore entirely possible to have IT professionals looking for work while employers simultaneously claim that they cannot find the skills they need.
What we are facing may no longer be a shortage of people. It may be a shortage of people who precisely match a particular requirement, can be productive immediately and happen to be available at exactly the point when a company wants to buy their expertise.
Perhaps we have also stopped creating the experts we are looking for
This raises another problem.
Before you can become senior, someone has to give you the opportunity not to be senior.
An experienced programme manager did not wake up one morning with fifteen complex migrations behind them. At some point, somebody entrusted them with their first one. An architect first had to take part in architecture decisions. A security specialist learned by working on imperfect systems under the guidance of more experienced colleagues.
When organisations systematically buy expertise whenever a problem becomes difficult, they also remove opportunities to develop that expertise internally.
The result is an odd situation: everyone wants senior professionals who can hit the ground running, while fewer organisations are willing to accept the time and risk involved in developing them.
The skills shortage then becomes, at least in part, self-inflicted.
IT may simply be ahead of the curve
This would already be an interesting development if it were confined to IT. Artificial intelligence gives it a much broader significance.
IT has already been through outsourcing, offshoring, SaaS, cloud computing, managed services and automation. In that sense, it has had a head start in turning permanent in-house roles into capabilities that can be purchased when required.
AI may now bring a similar logic to a much larger part of the service economy.
The usual question is whether AI will replace lawyers, accountants, developers, analysts or consultants.
That may be the wrong question.
A profession does not have to disappear for its labour market to be fundamentally disrupted.
A company may simply need 300 hours of specialist work where it previously needed 1,000.
The lawyer remains necessary. So does the accountant. The architect, developer and analyst do too. But fewer human hours are required to produce the same outcome.
That means demand for a particular skill can remain high while demand for the amount of human labour associated with that skill falls.
This distinction matters: a shortage of skills does not necessarily imply a shortage of full-time jobs.
Fractional roles may become far more common
One natural consequence could be the growth of fractional roles.
Two SMEs may not each need a full-time CISO. They may, however, each need 20 or 30 per cent of an excellent one. The same could apply to CTOs, architects, specialist lawyers, data leaders or programme managers.
There is a real benefit to this model: organisations that could never justify hiring such people full time gain access to expertise that would otherwise be beyond their reach.
But we should not pretend that this solves the employment equation.
If two full-time jobs are replaced by one specialist working 50 per cent for each company, fractional work has not created two jobs. It has simply distributed one full-time equivalent more efficiently.
The question of total demand for human labour remains.
Perhaps we need to look beyond the service economy
Over time, our societies have developed a rather peculiar hierarchy of work.
Knowledge-intensive service jobs are widely associated with professional success. Many jobs in agriculture, manufacturing and the skilled trades carry less prestige. Care work is recognised as essential while often remaining relatively poorly paid. Activities involving creativity, culture, sport, community or human attention are still easily dismissed as hobbies unless they generate substantial income.
Yet there is nothing inevitable about this hierarchy.
For several decades, we may have been confusing intellectual work with work that is difficult to replace.
Writing a summary, finding information, drafting a document, producing routine code or preparing a presentation used to require both human expertise and a significant amount of human time. Some of those capabilities are rapidly becoming less scarce.
Meanwhile, running cables through an old building, repairing an installation, maintaining machinery or working on physical infrastructure remains remarkably difficult to automate at a competitive cost.
It is not enough for a robot to be technically capable of performing a task. It has to perform it in a messy, variable, real-world environment — and do so more cheaply than a skilled human.
Care work raises a different issue. We can automate its administration, improve diagnostics and give professionals better tools. But caring for an elderly person, looking after a child or supporting someone who is dependent on others involves presence and human connection that cannot easily be reduced to information processing.
Perhaps the problem is not that these activities lack value. It is that we have historically been poor at turning that value into pay.
Diversifying income rather than searching for the one lifelong career
If demand for labour becomes increasingly fragmented, the way individuals organise their working lives may change as well.
In some economies, having several occupations is already commonplace. Someone may run a small shop, repair bicycles or motorcycles, rent out a room, sell goods at a market and provide other services on the side. Household income does not necessarily depend on a single occupation.
There is no reason to romanticise this model. It is often driven by necessity and can come with precarious income, very long working hours and limited social protection.
But the underlying principle deserves attention: do not make your entire income dependent on a single activity whose market value may change rapidly.
In wealthier economies, this could take a different form: part-time employment combined with freelance work, occasional specialist assignments, a small craft or commercial activity, and perhaps income from a product or another asset.
Instead of having one job, people may increasingly have a portfolio of activities.
But diversification cannot become another obligation
There is a substantial personal cost to this model.
Managing several activities requires time, energy, reasonably good health, the financial capacity to absorb fluctuations in income and, often, a professional network. Not everyone has those resources.
Our needs also change throughout our lives.
Young parents may quite reasonably prefer predictable employment for several years. The same is true of someone dealing with health problems, significant caring responsibilities or simply very little financial room for uncertainty.
Stable employment therefore does not need to disappear.
It may simply cease to be regarded as the only desirable end point of a successful career. At some stages of life, security will matter most. At others, the same person may prefer greater autonomy and several sources of income.
The challenge would then be to make it easier to move between these models rather than treating the secure employee and the independent entrepreneur as two opposing identities.
Skills cannot be reallocated like hours in a spreadsheet
There is one final and significant difficulty.
If some service-sector activities require less human labour in the future, we cannot simply move people into whichever sectors happen to be growing.
Skills are not interchangeable.
Many technical, industrial, craft and care professions depend on procedural knowledge developed through practice. Physical skill, observation, automatic responses and the ability to sense that something is simply “not right” can be the product of thousands of hours of experience.
An experienced analyst cannot be turned into an excellent mechanic, electrician, technician or care professional in six months.
Diversification has limits for the same reason. Some skills can be used occasionally without deteriorating significantly. Others require regular practice to maintain the gestures, reflexes and tacit knowledge sometimes loosely described as muscle memory.
A portfolio consisting of four entirely different occupations at 25 per cent each is therefore unlikely to be a universal model. Many people may instead retain one core occupation, with other skills and sources of income built around it.
This may be a question of value before it is a question of technology
That may ultimately be the harder challenge.
There is already extensive debate about how the productivity gains from automation should be distributed. It is an important question. If the benefits of automation become excessively concentrated, the economic and social consequences will be difficult to ignore.
But financial redistribution alone may not be enough.
If we continue to assume that abstract knowledge work naturally sits at the top of the professional hierarchy, that manual work is something people fall back on, that care is almost a vocation rather than skilled work, and that creative or attention-based activities are merely entertainment, we will preserve a hierarchy of value that may increasingly diverge from the actual scarcity of human capabilities.
AI could therefore have a rather paradoxical effect.
It may not eliminate work so much as change what we consider scarce — and therefore valuable.
IT may already be giving us a glimpse of that future: a skill can remain essential even when having someone with that skill permanently on the payroll no longer is.
If the same phenomenon spreads across the wider service economy, the question will no longer simply be which jobs AI will destroy and which new ones it will create.
It will be how we choose to distribute work, security, income and recognition across forms of activity whose relative value is changing.
And on that front, technology is likely to move much faster than our assumptions about work.
To discuss this: get in touch.