Where Juniors Start When AI Does the Simple Work

A colleague of mine recently raised a question which I keep hearing: if AI takes over the simple tasks which seniors used to hand over to juniors, what do juniors learn the job with, and why should we hire them at all? For a long time, these tasks were the way into the job, e.g. a small bug fix, the first draft of a test, the summary of a long document or the research for a presentation. (A senior could have done most of them faster, but they were a good way to learn.) If an assistant does them in a few seconds, this way into the job disappears, and so far I haven’t heard many answers to the question.

There are early signs that this already affects hiring. In August 2025, Erik Brynjolfsson, Bharat Chandar and Ruyu Chen of the Stanford Digital Economy Lab published an analysis of US payroll data. For workers aged 22 to 25 in the occupations most exposed to AI, such as software developers and customer service representatives, it showed a relative decline in employment of 13 percent since generative AI came into wide use in late 2022, while the employment of older workers in the same occupations stayed stable or grew. The authors point out that other factors may play a role, too.

In this article I would like to draw a picture of how people at the start of their career can enter the job today and how we as companies can deal with it. I’ll begin with how I started my own working life in the 1990s and whether this model works for AI. Then we look at the reasons for giving this task to juniors and at what to look for when we hire them.

How My Career Started

When the internet gained momentum in Germany in 1993, I was studying law. My university had a small computer room with UNIX terminals, which had no graphical user interface, and that is where we checked our emails. Hardly anyone outside universities and research institutes had access to the internet at that time. (In the same year, CERN put the software of the World Wide Web into the public domain, and the first graphical web browsers such as NCSA Mosaic came out, so a lot was about to change.)

I was quickly fascinated by what the new technology made possible, so I learned a lot about it besides my studies, and I set up and looked after networks and client computers for companies. After a while I felt quite confident with it.

When I started working, I knew much more about the new technology than most of my clients’ employees, as I had put my focus on it. The companies wanted to find out what it could do for them, so they brought me in to work out together with their employees where IT could improve their work. While doing this, I learned how their business worked from the people who had been doing it for years, and they learned how IT could support it.

Does the Same Model Work for AI?

I think it does. If we hire people at the start of their career who have put their focus on the new technology, they can find out together with our experienced employees where AI creates value, and they can build the solutions together. As they have not learned the existing ways of working yet, they have nothing to unlearn, and they look at each process with the question of what the new technology could change about it.

Many families know this pattern. When grandparents get a new smartphone, the grandchildren often set it up, while the grandparents know what they want to use it for: whom they want to call and which photos they want to see. While they set it up together, the grandchildren learn what matters to their grandparents, and the grandparents learn what the phone can do.

In a company the same thing happens at a larger scale. Let’s say the purchasing department of a manufacturer has twelve people who have negotiated with suppliers for many years. They know which suppliers deliver on time and which contract clauses have caused trouble before. A junior joins the team who has spent the last two years building small applications with language models. Together they look at where the department spends its time, and they find out that checking incoming offers and contracts against the company’s standard terms takes about a day per week. The junior builds a first assistant for this check, the purchasers review its results and explain why one clause matters and another one doesn’t, and after a few months the junior knows a lot about purchasing, while the purchasers know what an assistant can do and where it needs their judgment. (The details of this example are constructed, while the pattern is the same as when I started.)

The pattern looks like this:

How Juniors Learn the Job

This answers the first half of the question. In this model, juniors learn the job while they work with the experienced colleagues on the question of where AI helps. Every use case takes them into another part of the business, and the people who know this part explain to them how it works and why it works that way. The simple tasks of the past were one way into the domain, and the joint work on AI is another one, which may even be faster, as the junior contributes something useful within the first few weeks.

It also answers the second half: we hire juniors to explore together with our employees where AI creates value, and a few years later we have people who know both the domain and the new technology.

This leads me to a thesis: people at the start of their career have always done well to put their focus on what’s new and to learn the existing things step by step, as far as they don’t become obsolete in the new world. Some of today’s work will disappear with AI, and a junior who learns the existing work along the way mostly learns the parts which stay. This is how I started, and I’d expect it to work for AI, too.

Why People at the Start of Their Career?

You might wonder why we should give this task to juniors and not to experienced people who know AI well. There are many reasons, and these three are the most important to me:

1. On average, people at the start of their career have less reason to keep things as they are. Someone who has shaped a process for fifteen years knows its reasons and its weak points, and looks at a change with the caution this knowledge brings. A career starter does not have a way of working of their own to protect yet. They want to leave their mark and create something new, and they bring curiosity and a lot of energy. They aren’t afraid of change either, as the start of a career is a big transformation anyway.
2. They come into an environment with a lot of knowledge about the domain and its processes, which can be quite impressive. Many are grateful when they can work on interesting things at a well-known company for the first time. The experienced employees feel good, as their knowledge is admired and respected, while the juniors are happy that they can learn from them. This can become a win-win situation if both sides are open to it.
3. People at the start of their career usually ask for a lower salary than seniors. With a new technology, we’re at the very beginning and do not know yet where the value will be, so we keep the risk lower if the people who explore it aren’t too expensive.

These are averages, of course. There are curious seniors and cautious juniors, and the reasons above are about the career stage: someone who switches to a new field at 40 is at the start of a career, too.

Can’t Our Employees Ask the AI Themselves?

There is an obvious objection: do we need juniors to use AI, or can our employees ask the AI directly? The latter is possible, and many will do it. Yet, people at the start of their career are on average more curious and faster with new technologies, and bringing them in can speed up the whole change.

Companies have used this idea before. In 1999, Jack Welch, then CEO of General Electric, had his top managers pair up with younger employees who taught them how to use the internet. Most sources say that about 500 managers took part, and the approach became known as reverse mentoring. In the model above, the knowledge flows in both directions: the juniors learn the domain, and the experienced employees learn the new technology.

What to Look For When Hiring

My plea is this: when you staff your teams, make sure to bring in career starters who can show their curiosity, or test their curiosity and their willingness to work things out on their own. Two things can help with this:

• Ask what they’ve taught themselves recently and what they built with it.
• Give them a small task from a domain they don’t know yet, allow them to use AI, and talk with them afterwards about what they learned and how they checked the results.

Once they’re on board, put them in a team with experienced colleagues and give them a real question from the business, e.g. where AI could save the purchasing department a day per week. In most cases that’s a better start than a list of simple tasks which an assistant could do as well.

Assumptions

I assume that AI changes work in a similar way to the internet, so that companies need several years to find out where it creates value. If AI turns out to be a tool which everybody learns in an afternoon, the head start of the juniors will be smaller. This article also does not cover how juniors learn the craft of their profession, e.g. how to write maintainable code or how to judge the result of an AI. They need this, too, and they’ll learn it best from experienced colleagues, but that’s a topic of its own.

Conclusion

When AI takes over the simple tasks with which juniors used to learn the job, juniors can start where I started in the 1990s: with the new technology. Companies can hire them to find out together with their experienced employees where AI creates value and to build the solutions, and along the way the juniors learn the domain while the experienced employees learn what AI can do. Juniors usually cost less while the value is still unclear and they’re open to change, and when both sides respect each other’s knowledge, the collaboration can become a win-win situation. So when you hire, look for people at the start of their career who can show their curiosity and their willingness to work things out on their own.