If AI takes the junior work, who develops the next senior team?

A conversation that raised a harder question

A CTO at a large, established organization told me they had stopped hiring junior engineers because AI could now do the work. The immediate implication was striking. The longer-term implication was more interesting.

He still expected the organization to need senior engineers years later. If junior roles disappeared, though, how would the next generation acquire the experience those senior positions required?

That was one person’s account of one organization. It was not a labor-market study. But it exposed an operating question that reaches beyond engineering.

Production and development happen in the same work

An entry-level task can have two purposes. It produces something the business needs, and it helps a person develop judgment through repetition, feedback, and increasing responsibility.

If technology reduces the cost of producing the output, the learning function does not automatically reappear somewhere else. A team can improve this quarter’s efficiency while quietly weakening the path by which people become more capable.

That possibility matters for agencies because the quality of service depends on people making decisions in context. Knowing how to produce an output is only part of knowing whether it is the right output for the client.

The risk is a mismatch, not a single forecast

One possible outcome is a shortage of experienced people alongside fewer entry-level opportunities. Another is that AI changes senior roles as substantially as junior ones. The original conversation did not settle that question, and this article does not claim to predict it.

The useful takeaway is to make the dependency visible. If the future operating model requires expert judgment, the business needs a credible explanation for how it will develop or obtain that judgment.

Ask what experience the work is supposed to build

For a service team, a practical starting point is to identify the decisions a developing employee needs to learn. Which tasks teach them to spot a weak assumption, recognize a client constraint, or evaluate the quality of a recommendation?

When AI changes those tasks, change the learning design intentionally. A review of an AI-assisted output can still be substantive if the employee has to explain the tradeoffs, verify the evidence, and connect the recommendation to the client’s objective.

The practical question for agency leaders: do not measure a role only by the volume of output it produces today. Also consider the capability it is supposed to develop. For an account leader, that could mean learning to identify a client-retention risk, investigate it, and choose a response the client can trust.

Keep the standard visible

A clear scorecard helps make development concrete. The same framework that defines expectations during hiring can show an employee which skills are improving and which need more practice.

VantaFive’s People Evaluation System connects initial assessment to recurring reviews for that reason. AI may change how much work a person can do, but teams still need a consistent way to discuss the quality of the work and the judgment behind it.

The question is not whether every old task or role should be preserved. It is whether a new operating model includes a deliberate path for people to become excellent.