There is a management paradox rarely called by its name, even though almost every growing team faces it: the better a person is at doing specific hands-on work, the harder it is for them to stop doing it. In management literature, this is known as the competence trap — a concept developed by organizational learning researchers, including Barbara Levitt and James March in the 1980s.
The mechanics are simple and insidious because of it. An employee handles a task well — faster and with higher quality than colleagues. The team and the manager take notice and, quite logically, assign these exact tasks to this person more often than to others. The employee continues to improve at them — becoming even faster and more accurate. The circle closes: the more competent they are at a specific operational task, the less physical time and organizational reason they have to learn something else — such as delegating, mentoring, or designing a process rather than executing it.
The trap works imperceptibly because each individual step looks rational: both for the employee and for the manager. Nobody deliberately decides "let's bury a strong specialist in routine forever" — the decision is made time after time, episode after episode, each time optimal in the moment, yet cumulatively leading straight there.
People considered "reliable" and "the ones who will figure it out fastest" are especially vulnerable to this trap — seemingly a compliment, but in practice often meaning that this exact person won't be given the room to transition to higher-level tasks.
What helps in practice:
deliberately allocating time for growth tasks to strong performers — not "when there's a free minute," but protected hours on the calendar;
shifting the criterion for praise from "did it quickly myself" to "taught someone else how to do it";
when distributing tasks, sometimes consciously assigning the routine not to the fastest person, but to the one who still needs to master it.