How Michael Emery Thinks About Training, Coaching, and Mentoring in Safety

Episode 14 of the Headline Podcast, published on February 11, 2026, features Michael Emery, founder and director of Securus Health & Safety. Emery's central argument is that coaching is a questioning and listening exercise, which changes the safety professional's task from correcting visible behavior to understanding the decision that produced it.

That distinction matters because a conversation can look respectful while still functioning as a warning. The person leaves with a reminder, the form is closed, and the same exposure returns during the next busy shift. The stronger test is whether the discussion reveals a constraint that leaders can remove and a judgment that the worker can strengthen.

Coaching is not a softer compliance check

Coaching is useful when it improves the quality of a future decision. It is not useful when the coach already has the answer, asks a few questions for appearance, and then delivers the same instruction that a compliance officer would have given at the start.

Emery's position is precise because it protects coaching from becoming a tone change. A firm conversation can still be coaching, while a warm conversation can still be a disguised reprimand. The difference is whether the safety professional is trying to understand the work before deciding what kind of help is needed.

That approach aligns with James Reason's account of latent organizational conditions. A deviation may be the visible part of a chain that includes unclear priorities, poor access, production pressure, weak handover, or a control that is technically present but difficult to use. The conversation should test those conditions instead of assuming that intention explains everything.

OSHA's guidance on safety and health programs places worker participation alongside management leadership and hazard identification. That sequence matters because listening is not a courtesy added after the control review. It is one of the ways a leader discovers whether the control can work in the conditions where people actually use it.

The first question should locate the decision

A productive conversation starts by locating the decision in time and context. Ask what the worker saw, what changed, what outcome mattered in that moment, and which safer option was available without pretending that the work was static.

“Why did you do that?” invites a defense of character. “What was happening when you made that choice?” invites evidence. The second question does not excuse a hazardous action, but it gives the supervisor a better chance of finding the trade-off that made the action seem reasonable at the time.

A supervisor can structure the first five minutes around four facts. Identify the task state, the information available, the control that was expected, and the consequence that the worker was managing. If the account changes when those facts are discussed, the change itself may reveal a communication or trust problem.

ISO 45001:2018 emphasizes worker consultation and participation because people closest to the work hold information that formal procedures may miss. The conversation therefore has a dual purpose. It develops individual judgment while testing whether the management system has made the intended decision visible and practical.

Training, coaching, and mentoring solve different problems

Training addresses a defined knowledge or skill gap, coaching examines a live decision, and mentoring supports longer-term development. Treating all three as the same intervention produces vague actions and makes it difficult to tell whether the follow-up changed the risk.

Suppose a worker has never been shown how to inspect a lifting accessory. The immediate response is training, followed by a demonstration and a verification step. If the worker knows the inspection method but skips it when the schedule compresses, the issue may require coaching about decision pressure and a management response to the schedule.

Mentoring has a different time horizon. A junior safety professional may need repeated reflection on how to frame questions, read evidence, and influence a supervisor. That development cannot be completed in one field correction, because the skill is built through practice across several situations.

The table prevents a common failure. When a supervisor labels every issue as a training need, the organization can keep sending people to courses while leaving the work design untouched. Emery's coaching lens forces the team to ask which response matches the actual problem.

Listening reveals the constraint behind the behavior

Listening is not passive agreement. It is an evidence-gathering discipline that tests whether the worker faced a time, tool, information, coordination, or authority constraint that altered the decision.

The safety professional should listen for the nouns of the situation, not only the verbs of the action. “I climbed” describes behavior. “The access platform was blocked, the permit expired, and the isolation owner was unavailable” describes conditions that may require leadership action.

A useful follow-up has three layers. First, ask the person to describe the sequence without interruption. Second, repeat the critical facts and check whether the summary is accurate. Third, ask what change would make the preferred action more achievable during the next shift.

The final question is where coaching either becomes operational or remains conversational. If the answer is “people need to be more careful,” the conversation has not yet found a control. If the answer names a staging change, a handover rule, a decision right, or a clearer escalation route, the team has something it can test.

The HSE's approach to worker involvement similarly treats participation as a route to better risk control, not simply as a morale exercise. Listening has value when it changes the hazard picture, the control design, or the leader's response time.

Compliance policing changes what workers disclose

When workers expect a safety professional to act as a compliance policeman, they optimize for a clean interaction rather than an accurate account. The organization then receives polished answers while losing the conditions that explain why work drifts from the procedure.

That does not mean standards or accountability should disappear. A known breach may still require a clear boundary, a stop-work decision, or formal action. Coaching improves the quality of that response by separating the immediate risk decision from the later question of why the decision became attractive.

Andreza Araujo's work on safety culture makes the same distinction between visible compliance and operating culture. A signed form, completed observation, or closed action can show that a process occurred without proving that people trust the process enough to disclose a constraint.

Psychological safety is relevant here because workers need a credible route for raising bad news. Amy Edmondson's research frames that climate as a condition for speaking up, although speaking up alone is not a control. Leaders must respond with action, because repeated requests for information without visible change teach people that disclosure carries effort without benefit.

Leaders should measure the next decision, not the conversation count

The most useful measure is whether coaching changes what happens next. Leaders should review the quality of the evidence gathered, the speed of the system response, repeat deviations, and whether workers raise constraints before exposure becomes urgent.

Counting conversations is easy, which is why it can become a misleading target. A supervisor can complete 20 interactions in a week and still fail to identify one recurring access problem. A smaller number of detailed conversations may be more valuable if they produce a change that holds during the next 7 or 14 days.

Use a simple review sheet with five fields: decision observed, condition reported, intervention selected, owner named, and verification date. After 14 days, check whether the action changed the work, whether the behavior returned, and whether the worker's explanation became more specific. The pattern tells leaders whether coaching is improving judgment or merely improving documentation.

Link the review to existing indicators rather than creating a new dashboard. A rise in near-miss reporting can mean trust is improving, but it can also mean exposure increased. Pair the count with field verification, response time, and evidence that controls changed. The point is to improve decision quality, not to create a new number that looks impressive.

For a related framework, compare this approach with the five moments when coaching becomes a control failure, the evidence levels for observation quality, and the proof points for training transfer.

Recommendation

Start with one recurring deviation and test whether a coaching conversation can expose the constraint, assign a system response, and improve the next decision within 14 days. Keep formal accountability where the risk requires it, but do not confuse a completed correction with a changed condition.

Michael Emery's argument is valuable because it rejects a false choice between standards and listening. A safety professional can hold a clear boundary while remaining curious about the conditions that shaped the decision. That combination is more demanding than a reminder, yet it gives the organization a better chance of preventing repetition.

The practical sequence is short enough to use on the next shift. Listen to the account, locate the decision, identify the constraint, choose training or coaching deliberately, name the owner of any system change, and verify the result on a defined date. If the same behavior returns, treat the recurrence as evidence that the first response was incomplete.

As Andreza Araujo's book Safety Culture: From Theory to Practice argues, culture becomes visible through the decisions leaders reinforce. A coaching conversation becomes credible when it changes the work, protects honest information, and gives people a reason to raise the next problem earlier.

Listen to the full conversation with Michael Emery on the Headline Podcast, and compare the episode's coaching perspective with the OSHA safety and health program guidance, the ISO 45001:2018 framework, and the HSE worker involvement guidance.