A mechanical track includes a raised blue bypass that avoids a section of the original route.

When to stop developing a skill and redesign the work

To choose between developing a skill and redesigning the work, first locate the demand. It may lie with the person, an avoidable task, an interface, or variability in the system.

PRYSMAP5 min read

To compare the responses, start with what each addresses. Training allows you to assign participants, content, and dates. Redesign changes technology, workflow, authority, or the distribution of tasks. That is why it often looks more expensive, even when the gap keeps returning.

The decision can be structured as a sequence of questions. None produces an automatic answer. Together, they prevent every solution from defaulting to a change in the individual.

First question: should this demand exist?

Before developing the capability to perform a task, ask what outcome it protects. Duplicate controls, transcription, approvals, and reconciliations can outlive the conditions that created them. If an activity does not reduce a risk or create current value, eliminate it.

Do not confuse frequent use with necessity. A task may take many hours because the process is poorly designed. Nor should you eliminate a rare activity without analyzing the consequences of its absence. The goal is to remove unnecessary demand, not pursue superficial efficiency.

If the outcome is still needed, move to the next question.

Second question: is the barrier in the person or the system?

Data can vary and instructions can conflict. Permissions may arrive late, and tools may hide context. These conditions produce errors that look like individual gaps. Training people to compensate can improve their adaptation while leaving the source untouched.

Observe where the failure occurs. Check what information was available and whether competent people struggle at the same point. When the pattern cuts across experience levels, redesign takes priority.

ISO 6385 sets out ergonomic principles for designing work systems that consider interactions between people and technical or organizational components. It does not prescribe a particular talent decision; it supports a perspective that does not separate individual performance from the system in which it occurs. ISO 6385:2016 (opens in a new tab).

Third question: should you control the source before training the response?

NIOSH orders safety controls from elimination and substitution through engineering and administrative controls to personal protective equipment. Its purpose is occupational risk management, not general skills design. The logic can be transferred cautiously: addressing the source is often more reliable than relying exclusively on repeated individual behavior. NIOSH (opens in a new tab).

In a talent process, eliminating a manual burden or blocking an invalid option may be more effective than reminding everyone about the error. Making a critical signal visible serves a similar purpose. Training remains necessary when human judgment is an essential part of the control, but it should not compensate indefinitely for a design that can be fixed.

A blue lift table adjusts the height of a load to connect it with two different work surfaces.

Fourth question: what does it cost to sustain proficiency?

A rarely used skill may require ongoing practice to maintain. Compare that cost with frequency, consequence, and alternative ways to provide coverage. It may be better to concentrate the capability in a group, use a tool, or establish an external route.

Do not assess only the cost of the course. Consider learning time and opportunities to practice. Add supervision, early errors, and expected obsolescence. Obsolescence can change demand for a capability without making the person obsolete. A skill that can technically be taught may not be a reasonable investment for every role.

Fifth question: is the human response consistent enough?

Some decisions require sensitivity to context and allow reasoned variation. Others demand extreme consistency, high volume, or a faster response than a person can sustain. If small lapses have serious consequences and the environment allows a more reliable control, redesign the work.

Automation does not eliminate responsibility. It can shift that responsibility to people who define rules, verify data, monitor exceptions, and decide when to stop the system. The redesign must describe those new capabilities rather than present technology as a substitute that requires no human work.

When a decision has serious consequences and the system is uncertain, retain meaningful human review. A “human in the loop” is of no use if that person lacks the time, context, or authority to challenge the recommendation.

Choose among four responses, not two

At the end of the diagnosis, there are four ways forward. You can develop, redesign, combine, or stop doing the work. Develop when the capability is intrinsic to the outcome and practice can improve it. Redesign when the demand or barrier lies mainly in the system. Combine when a better environment and new human judgment depend on each other. Eliminate when the outcome no longer justifies the work.

Document which evidence led to the choice and which signal would require you to revisit it. A redesign can create new exclusions. Training may not transfer to practice. Automation can degrade. A responsible decision includes follow-up.

Measure capability, not learning activity

After the intervention, observe quality, time, workload, recovery, and the distribution of errors. If training increased but the bottleneck remains, do not conclude that people need even more of the same content.

Also review who absorbed the displaced work. A simpler interface for one team may create manual review for another. Redesign increases capability only if it improves the whole system rather than hiding its costs.

Stopping the development of a skill does not mean stopping people’s development. It can free time for more transferable capabilities, higher-value decisions, or less fragile ways of working. Decide based on the combination of person and system that produces the outcome sustainably, not on how much more someone can learn to endure the current design.