If AI replaces junior staff, where will tomorrow’s seniors come from? And how will we train them?

According to SignalFire’s ‘State of Tech Talent 2025’ report, hiring of new graduates by major tech companies has dropped by more than 50% over the last five years.

Article published in “Via Empresa”, June 2026

By Oriol Amat

Economist. Professor at Pompeu Fabra University (UPF), the BSM-Barcelona School of Management, and the SME Observatory. He chairs the SME Advisory Council of Catalonia.

Artificial intelligence is already transforming the labor market. For the moment, we are not witnessing mass job destruction. What *is* being observed is a significant reduction in the hiring of younger professionals and a profound shift in the requirements companies demand of new candidates.

The initial evidence is quite compelling. According to SignalFire’s *State of Tech Talent 2025* report, the hiring of new graduates by major tech companies has fallen by more than 50% over the past five years. At the same time, in countries like the United States, the unemployment rate among recent university graduates has risen by 30% since 2022. Conversely, companies are increasing their demand for mid-level and senior professionals. Along these same lines, PwC’s *AI Jobs Barometer 2026* concludes that organizations are cutting back on onboarding entry-level talent while simultaneously requiring new professionals to possess competencies that were, until recently, reserved for those with more experience—such as judgment, leadership, emotional intelligence, and decision-making in complex environments.

In United States, the unemployment rate among recent university graduates has risen by 30% since 2022.

What is happening? For decades, companies have developed their talent through a gradual learning process. Professionals joining the workforce used to start out by performing relatively simple tasks—searching for information, drafting documents, analysing basic data, reviewing reports, or preparing presentations—and, as they gained experience, they took on increasingly significant responsibilities.

Many of the tasks that traditionally served as a training ground for professionals to learn the ropes can now be performed by artificial intelligence tools in a matter of seconds. Consequently, companies require fewer people to carry out these functions and tend to focus their hiring on candidates with more experience or higher-value skills. In a sense, AI is eliminating some of the initial rungs on the career ladder.

From a business perspective, this evolution is perfectly understandable. Organizations have always adopted innovations that boost productivity, improve quality, and allow them to achieve more with the same resources. This happened with computing, then with the internet, and is now happening with artificial intelligence.

Transformation and riscs

This transformation raises some important questions.

• The first concerns young people’s access to the labour market. If companies reduce entry-level opportunities, many graduates may struggle to gain the experience that will later be required of them. There is a risk of falling into a paradoxical cycle: you need experience to get a job, but you need an initial job opportunity to gain experience.
• The second issue is the risk of widening inequality among professionals. Individuals capable of using artificial intelligence to boost their productivity and deliver greater value will be increasingly in demand. Conversely, those who lack the necessary skills or struggle to adapt to new work environments may see their career opportunities diminish.

However, the most significant challenge likely lies elsewhere, directly affecting companies’ ability to secure the talent they will need in the future.

If fewer junior staff are hired today, where will tomorrow’s senior professionals come from? This is one of the major paradoxes AI introduces to the labor market.

Companies need managers, project leaders, consultants, auditors, lawyers, engineers, and many other professionals with experience and sound judgment. However, these profiles are built over the years through practice, successes, mistakes, and the accumulation of knowledge.

The history of any organization is full of people who started in modest positions and, thanks to the experience they gained, eventually took on key responsibilities. For this reason, a sustained reduction in the hiring of entry-level talent could lead to an unexpected consequence: a diminished pool of available talent.

From medieval guilds to today’s large corporations, every profession has relied on apprentices.

The very process that drives efficiency today could end up weakening the pipeline of professionals upon which many organizations’ future growth depends. This consideration is particularly relevant in a context defined by an aging population, low birth rates, and growing difficulties in finding qualified professionals across numerous sectors. If opportunities for young people to gain professional experience are also excessively curtailed, the problem could intensify in the coming years.

The solution is not to halt artificial intelligence, but to harness it. Educational institutions will need to fully integrate it into their curricula; companies will have to redesign career paths so that young people can learn faster and take on higher-value responsibilities; and professionals will need to incorporate these tools into their daily work. At the same time, it will be necessary to strengthen capabilities that technology cannot easily replace: critical thinking, professional judgment, creativity, leadership, teamwork, empathy, accountability for one’s work, and decision-making in uncertain environments.

The good news is that AI does not eliminate the need for talent. We will continue to need people capable of making decisions, building trust, innovating, leading teams, and interpreting complex situations. What is changing is not the importance of talent, but the way it is developed. The question is no longer whether AI will replace professionals, but whether we can adapt our educational and corporate systems to cultivate the talent we will need.

AI is already changing many things. Yet one reality remains unchanged: the companies of the future will need experienced professionals, and such professionals will only exist if someone gives them their first opportunity.

How will we train tomorrow senior professionals?

The challenge is not to protect these jobs, but to reinvent learning.

The last conclusion leads to another question: how will we train senior people if AI is already performing many of the tasks through which they used to learn? For decades, learning a trade largely meant starting at the bottom. An auditor would review invoices; a lawyer would research case law. These were not always the most stimulating jobs, but they had a great virtue: one learned by doing. Now, AI can perform many of these tasks in seconds. We can eliminate the task, but we cannot eliminate the learning it provided.

1. Protect learning, not tasks

Image: iStock by Getty Images

The solution is not to artificially preserve old junior-level jobs, but to safeguard the process that allows for the acquisition of knowledge, experience, and judgment. A recent McKinsey study involving nearly 1,500 executives and talent managers indicates that one-third of companies are already reducing the basic tasks through which juniors traditionally learned. However, 66% predict that AI will actually increase the hiring of entry-level professionals. This will not happen everywhere; in some fields—especially where AI can take over a large portion of initial tasks—the need for juniors will indeed decrease.

Therefore, it is not simply a matter of hiring more or fewer people. We will need different kinds of juniors and, likely, in different proportions depending on the specific activity.

2. Use AI to learn, not to stop thinking

If juniors need to be different, the way they learn must also change. McKinsey proposes a simple formula: first the person thinks, then they ask the AI, and finally, they compare. Imagine a young analyst tasked with valuing a company. If they ask the AI ​​to perform the analysis directly, they will save time. But what will they have learned?

We can do it the other way around: first, study the company, identify the risks, and reach some conclusions; Then, ask the AI ​​a question and compare the answers with those of an experienced professional. What has the AI ​​seen that the professional hadn’t? And what does the professional know about the client, the industry, or the context that the AI ​​doesn’t?

And the process continues. It’s also necessary to learn how to ask and re-ask the AI ​​questions. An answer can open up new questions, suggest areas for further exploration, or contain conclusions that need to be challenged.

The key is to iterate: ask, compare, re-ask, and improve. This requires critical thinking.

A Microsoft Research study, based on 936 real-world examples of generative AI use, observed that the more professionals trusted the AI, the less critical thinking they reported applying; conversely, the more they trusted their own knowledge, the more they questioned, compared and contrasted the answers. AI can boost productivity but also reduce the mental effort required for thinking. However, when used effectively, it can facilitate faster learning and progress: it can become an extraordinary tool for accelerating the learning process.

3. Senior professionals must take on more of a teaching role

This new model also shifts the role of experienced professionals. If AI can explain a procedure, search for information, or prepare a draft, the value of the senior professional shifts toward what is hardest to automate: imparting judgment. Why does an operation that looks excellent on paper fail to convince us? In what cases might a deviation be the first sign of a problem? This knowledge is accumulated over years through trial and error.

Therefore, companies must give senior professionals time to train those coming up behind them and must recognize this function. Until now, we primarily rewarded the senior professional who solved problems effectively. Increasingly, we must value the senior professional who teaches others how to solve them.

Until now, we primarily rewarded the senior professional who solved problems effectively. Increasingly, we must value the senior professional who teaches others how to solve them.

Thus, AI does not eliminate the need to develop talent; it compels us to develop it differently. The more tasks AI can perform, the greater the value of what only people can contribute: exercising judgment, building trust, assuming responsibility, and helping others grow.

4. Seven actions to develop the senior professionals of tomorrow. What can companies do?

• Continue hiring junior professionals. Fewer may be needed in some areas, but we cannot shut off the talent pipeline and expect to find senior professionals when we need them.
• Expose them to real-world problems sooner. Less mechanical work and more client interaction, meetings, projects, and decision-making—always under supervision.
• Use AI for learning, not just for output. Think first, then ask, compare, and ask follow-up questions.
• Turn senior professionals into mentors. Experience only creates long-term value when it is passed on.
• Develop judgment. Technical knowledge will remain essential, but critical thinking, curiosity, communication, empathy, creativity, and the ability to make decisions based on incomplete information will become increasingly important.
• Conveying values ​​and ethical standards. Senior professionals must also help junior staff act with integrity.
• Learning to take responsibility. AI can assist with analysis and decision-making, but it cannot assume responsibility for the outcome. It is necessary to verify answers, acknowledge errors, and stand behind the work one signs or presents.

AI allows us to learn faster, which presents a tremendous opportunity. However, accelerating learning does not mean eliminating it. We may not be witnessing the disappearance of the apprentice, but rather the reinvention of learning. AI can shorten the path from junior to senior, yet we must still walk that path ourselves.