For a long time, the way into most careers has followed a similar pattern. You start in an entry-level role, learn by doing, pick things up from the people around you, and gradually build experience. It’s not always efficient, and it’s rarely glamorous, but it creates a path.

That path is starting to shift.

AI is increasingly capable of handling the kinds of tasks that often sit at the entry point. Drafting emails, summarising information, handling basic customer queries, organising data. The sort of work that used to be repetitive, but useful as a way to learn how things operate.

On the surface, this looks like progress. If those tasks can be done faster and more consistently, it frees people up to focus on more complex work. But it also raises a quieter question about what happens to the roles that used to sit underneath that complexity.

The role entry-level work used to play

Entry-level roles have never just been about the tasks themselves. They’ve been a way of understanding context. How decisions get made, how systems connect, how people handle situations that don’t quite fit the process.

A lot of that learning isn’t formal. It comes from being around the work, noticing patterns, making mistakes, and gradually building judgment.

When AI starts to take on the more structured parts of those roles, it doesn’t just remove tasks. It risks removing some of the exposure that helps people build that understanding in the first place.

Efficiency doesn’t automatically create opportunity

There’s an assumption that if AI removes certain types of work, new opportunities will naturally appear to replace them. And in some cases, they will.

But those opportunities don’t always look like the roles they replace.

If fewer people are needed to handle entry-level tasks, organisations may simply hire fewer people at that level. The work still gets done, just with a different structure. That can make it harder to find a starting point, particularly for people trying to enter an industry without prior experience.

It also shifts expectations. If the baseline tasks are handled by AI, the work that remains may require a higher level of judgment from the start.

Where AI academies start to come in

This is where ideas like AI academies and structured training programmes start to appear.

Instead of learning primarily through entry-level roles, there’s a move towards more intentional development. Teaching people how to work with AI tools, how to interpret outputs, and how to apply judgment in situations where the system doesn’t provide a clear answer.

In theory, this could accelerate learning. Rather than spending months or years on repetitive tasks, people are introduced earlier to the parts of the role that require decision-making and context.

But it also changes the nature of how people gain experience.

Learning becomes something that is designed, rather than something that emerges from being embedded in the work itself. That can be more efficient, but it may not fully replicate the messiness of real-world situations.

The risk of skipping a step

There’s a risk that in removing entry-level tasks, we also remove a layer of understanding.

If people are expected to make decisions earlier, but haven’t had the same level of exposure to how those decisions play out, there’s a gap. Not necessarily in knowledge, but in experience.

AI can support that to some extent. It can provide guidance, surface information, and suggest next steps. But it doesn’t fully replace the process of learning through doing.

A different kind of starting point

It’s possible that entry-level roles don’t disappear, but change.

Instead of being defined by repetitive tasks, they may become more focused on working alongside AI. Interpreting outputs, checking accuracy, handling exceptions, and understanding where systems fall short.

That would create a different kind of entry point. One that is less about doing the work itself, and more about managing how the work is done.

Where this leaves things

There isn’t a single answer to what replaces entry-level roles.

In some cases, they may reduce. In others, they may evolve. New forms of training, like AI academies, may take on some of the role that entry-level work used to play.

But the underlying question remains.

If experience used to come from doing the work, and the work is changing, how do people build that experience now?