看見「看不見」的訓練

From 10 a.m. to 12 p.m. on May 6, 2026, I visited the Special Search Brigade to share a competency assessment system for new personnel. The session started with a digital tool, but the discussion that resonated most was not about Google Sheets or Apps Script. It was about a question every training unit eventually faces: how do we know when a new member is truly ready?

Training new personnel is rarely a matter of not teaching enough. In frontline work, the harder problem is that instructors, mentors, and supervisors often teach, observe, remind, and correct many times, yet much of that process remains difficult to review later. A mentor may feel that someone is "almost ready" or "still unstable," but those judgments often stay as impressions rather than becoming shared training evidence.

For me, a competency assessment system is not simply a way to move paper forms online. It is a way to break training into visible parts:

• Which competencies must be confirmed before independent duty?
• Which observations should mentors record?
• Which feedback should trainees be able to understand and act on?
• Which delays, repeated errors, or unresolved gaps should supervisors be able to see?
• When competency does not develop as expected, should we first blame the person, or examine whether the training process is catching the risk?

The discussion mattered because the questions were practical. Participants asked about mentor handovers, documentation burden, what supervisors actually need to see, and whether a group of new members can be evaluated through a shared standard. These are the real questions that determine whether a training system can work outside a slide deck.

If a training system is detached from the field, it becomes only a polished form. The difficult work is to let the form carry the judgments, reminders, observations, and feedback already happening in practice. Competence does not appear suddenly on a single day. It is built through repeated observation, correction, successful performance, mistakes, and follow-up.

If all of that remains verbal, the organization cannot learn from it. If everything becomes rigid checkboxes, the field experiences it as another administrative burden. A good competency assessment design should sit between the two: it should give experience a place to be recorded, and make the record useful for training.

Translating Medical Education Into the Fire Service Context

The system was not only a form design project. It was an attempt to translate mature concepts from medical education into the realities of fire service shift work and new personnel training.

The first framework was Miller's Pyramid. In training, we often ask whether someone "can do it," but competence is not a single state. Knowing, knowing how, showing how, and performing in real situations are different levels. If training ends when a trainee nods and says they understand, but the final expectation is independent performance, the intermediate levels are left unverified.

The second issue is that self-assessment is unreliable. The Dunning-Kruger effect reminds us that people with limited competence may not be lying when they overestimate themselves. They may not yet have the ability to judge their own gaps. This is why new personnel training needs external observation, workplace confirmation, and repeated feedback.

The third concept is Workplace-Based Assessment. Fire service training does not happen with a single examiner in a classroom. It happens across shifts, calls, handovers, different mentors, and different real-world situations. In that environment, assessment should not rely only on one exam. It needs multiple observers, longitudinal accumulation, and records that return to the workplace. Many small observations, when accumulated, can form a more reliable map of competence.

The Hard Part Is Not Code

One point I emphasized in the session was that technology is not the hardest part. With AI assistance, writing Google Apps Script, creating forms, and building interfaces is much more accessible than before. The harder work is breaking frontline know-how into items that can be trained, observed, discussed, and improved.

A competency list is not about having an impressive number of items. The key is whether the decomposition is close enough to the actual work. It is not simply a list of broad domains such as EMS, rescue, driving, or station duties. It must be specific enough for mentors to know what they are teaching, for trainees to understand where they are stuck, and for supervisors to see where support is needed. That kind of decomposition requires field experience and a language for education design.

The core of the system is therefore not code. It is design logic: how to address inconsistent standards, invisible progress, and invisible teaching contributions; how to leave a record for those who are willing to teach; how to help supervisors see exceptions instead of drowning in data; and how to keep enough flexibility so the system does not collapse when reality creates exceptions.

This is why I continue to organize my work around EMS Safety and education design. Safety is not only about reminding people to be careful, and training is not finished when a class ends. The work worth doing is to take problems that frontline teams face every day, but often cannot easily name, and turn them into systems that can be trained, assessed, repeated, and improved.

If one session can help a team start asking, "How do we actually know someone is ready?", then it is not just a talk. It becomes a starting point for better training design.

If your unit is also thinking about new personnel training, mentor handovers, competency assessment, or EMS Safety education design, these questions are well suited for a workshop or focused training session. The tools can be copied, but the real work is to examine each unit's own operating structure.