IX Coach

Safety Leadership Coaching: What It Changes, and What It Cannot

What is safety leadership coaching, and does it improve safety performance?

Safety leadership coaching is coaching for the supervisors, foremen, site managers and executives who direct frontline work, aimed at the few behaviours that tell a crew what the real priority is. Here is what it works on, what the controlled studies actually found and where they stop, how it sits inside a high-hazard operation, how to judge a coach, and the problems it cannot reach.

Safety leadership coaching is coaching for the people who direct frontline work (supervisors, foremen, shift leads, site and plant managers, and the executives above them) aimed at the small set of behaviours that tell a crew what the real priority is: what the leader does in the minute after someone reports a problem, how often safety comes up when nothing has gone wrong, and which way the leader decides when the schedule and the safe way collide. It is different from safety training, which teaches people the rules, and from a safety management system, which writes the rules down.

Does it work? The controlled studies are few and small, but they mostly point one way: when leaders change a few specific behaviours and get feedback on them, the people who work for them change what they do. None of them reports fewer deaths, and none was run in a refinery or chemical plant, so treat any large promised reduction with suspicion. Below: what the coaching works on, what the research does and does not show, how it fits inside a high-hazard operation, how to judge a coach, and the problems it cannot reach.

What the coaching actually works on

The material is smaller and more behavioural than the label suggests. The first minute after a worker reports a near miss or a defect: does the leader ask what happened, or who did it. What the leader looks at, and asks about, on a walk through the operation. How a leader makes the call when finishing on time means a shortcut, and whether the crew can see that call being made. Whether safety is ever discussed on a day when nothing went wrong.

This matches how the US Occupational Safety and Health Administration frames the leader's part. Management leadership is the first of the seven core elements in its Recommended Practices for Safety and Health Programs, and its page on the element asks leaders to be visible in operations and to "Set an example through their own actions." It also asks them to set up ways for management and workers to talk about safety and health freely, without fear of retaliation, and says successful programs reward rather than discipline workers who identify problems. Each of those is a thing a person does on a particular Tuesday, which is why it can be coached.

A coaching engagement on this usually starts from what the crew currently believes will happen if they raise something, which is not always what the leader believes. It then works one situation at a time: a pre-shift talk, a stop-work decision, a conversation with a worker who took a shortcut, a debrief after something went wrong. A typical approach is to rehearse the conversation beforehand and review it afterwards.

Explore: the leader as coach · crucial conversations · effective one on ones

What the research shows, and where it stops

The strongest evidence is a handful of studies in which leaders were coached or trained and the result was measured. In a 2002 study in the Journal of Applied Psychology, Dov Zohar gave line supervisors weekly feedback on how often they were having safety-oriented interactions with their workers, and trained section managers to use the same information and to conduct follow-up interviews. Safety-oriented interaction rose significantly in the intervention groups and stayed flat in the control groups, and the rise was accompanied by lasting changes in minor-injury rate, earplug use and safety climate scores.

A 2010 study in the Journal of Safety Research tested the same idea on construction sites with foremen coached and given bi-weekly feedback on their daily safety conversations with their crews. At baseline, workers perceived safety as part of their foreman's talk in only 6 to 16 percent of exchanges. Safety communication rose significantly in the intervention gangs, and the observed level of safety rose by 7 percent and 12 percent in the two intervention groups, with no significant change in the three control gangs. The study covered five gangs in all, so it is a promising result, not a settled one.

A randomized study by Jane Mullen and Kevin Kelloway (2009) assigned 54 leaders from 21 long-term health care organizations to general transformational leadership training, safety-specific transformational leadership training, or a control group. Training produced significant effects on the leaders' own safety attitudes, intent to promote safety and self-efficacy, and on how employees rated their leader's safety-specific leadership and the safety climate. The abstract does not report a significant effect on injuries.

Two further sources widen the picture without adding a coaching trial. A 2002 study by Julian Barling and colleagues found that safety-specific transformational leadership predicted occupational injuries, through perceived safety climate, safety consciousness and safety-related events, in two samples (174 restaurant workers, then 164 young workers in diverse jobs). And Sharon Clarke's 2013 meta-analysis found that transformational leadership was associated with employees taking part in safety, while active transactional leadership was associated with both participation and compliance with rules. Her conclusion was that a combination of both styles appeared most beneficial. For a coach, that is a useful check on the idea that safety leadership is mainly about inspiring people: the active, follow-up side appears to matter alongside it.

Read these for what they are. Where the abstracts name a setting, it is restaurants, long-term care or construction sites, not refineries or chemical plants, and the Zohar abstract names none. Nothing summarized here reports a change in fatalities. If a provider quotes a percentage reduction in incidents, the useful question is which study it comes from, in which industry, and over what period.

Explore: transformational leadership

Coaching for safety performance: what to measure

The studies above measured the leader's behaviour itself, not just the injury count afterwards, and that is the practical lesson. In a single crew, serious events are rare enough that a quarter's tally says very little about whether a leader changed. How often a leader talks about safety with each worker, what happens after a report, and whether the crew sees a production-versus-safety call go the safe way are all observable within weeks.

So a coaching engagement worth paying for agrees up front on two or three observable behaviours, says who will observe them (ideally the crew, through short and regular check-ins, not only the leader's own account), and reviews them on a schedule. A leader's self-report that they are "more visible" is the weakest form of evidence on offer.

The same logic applies to near-miss reporting. A rise in reports after a leader changes how they respond can mean the crew trusts the process more, not that things got less safe, so a falling report count is not automatically good news.

Explore: after action review

In high-hazard industries, leaders sit inside a bigger system

"High hazard" is usually taken to mean operations where one failure can seriously hurt many people at once: chemicals and refining, oil and gas, mining, heavy construction, utilities. The consequence of a leader's call is larger there, which is why the rehearsal side of coaching matters: the stop-the-job decision, the unwelcome news passed up the line, the pressure to restart.

It is also where the limits of coaching a single leader are clearest. The US Chemical Safety Board's report on the March 2005 BP Texas City refinery explosion records that fifteen workers were killed and 180 others injured, and the board concluded that "Organizational and Safety Deficiencies at All Levels of the BP Corporation" caused the disaster. A finding of that kind is about the organization, and no amount of coaching a shift supervisor addresses what sits above and around them. A good safety leadership coach says which layer a problem lives in, including when the honest answer is that it is not the layer they are being paid to work on.

The pre-mortem, an exercise in which a team imagines the job has already gone wrong and works out why, is one tool a leader can use before a high-consequence task. It is a way of making room for the doubts a crew may not otherwise voice.

Explore: pre mortem · psychological safety teams

How to judge a safety leadership coach or program

Ask which behaviours they will work on and how those will be measured, by whom. A provider that can only offer a satisfaction survey is measuring the experience of being coached, not the leader's change.

Ask what they do if the honest finding is that the system, not the supervisor, is the problem: staffing, schedules, incentives that reward speed, a reporting process that punishes the person who reports. An answer exists, and a provider who has never considered it will run a communication skills program on a structural problem.

Ask whether they know your hazards. Sector experience matters here more than in most coaching, because a coach who cannot tell a routine deviation from a dangerous one cannot help a leader decide which conversation to have. Ask what they will not take on, which should include incident investigation findings, legal advice and compliance sign-off.

And ask for the evidence behind any number. The honest summary of the field is the one above: promising, small, mostly outside the highest-hazard settings.

Explore: the leader as coach

When coaching is not the right help

Coaching does not change a guard, a permit system, a staffing level or a procedure, and it does not replace the safety professionals, safety representatives or regulators whose job that is. If a hazard is unaddressed, the route is the employer's own safety function and, where that fails, the regulator in your country. If workers are being punished for raising problems by someone above the supervisor, that is a matter for senior management, a safety representative or a regulator, not a skills gap in the supervisor.

Serious injuries and deaths at work also leave people shaken: leaders who were on shift, crews who saw it, supervisors who wonder what they should have done. Shock, guilt and grief after an event like that are for the employer's support arrangements and for a clinician or counsellor, not for a coach, and nothing on this page is therapy, diagnosis or treatment. Coaching in general is not crisis care.

If you are having thoughts of suicide or of harming yourself, please contact a crisis line now. In the United States, call or text 988, available 24/7 and free. In Canada, call or text 9-8-8, 24 hours a day; if your safety is at risk, call 9-1-1. In the United Kingdom and Ireland, Samaritans is on 116 123, any time of day. In Australia, call Lifeline on 13 11 14 or text 0477 13 11 14; if life is in danger, call Triple Zero (000). Elsewhere, findahelpline.com lists helplines by country. If someone is in immediate danger, call your local emergency number.

Where IX Coach fits

For this work a human coach has real structural advantages. A human can walk the site, watch a pre-shift talk, interview the crew, learn the hazards of a particular plant, and work alongside a safety function that sponsors the engagement. An AI coach cannot do any of them.

IX Coach is an AI coaching product, not a human, not a safety adviser, not HR or legal, not a clinician and not a crisis service, and it does not detect crisis. What it is designed to be is a place for a leader to think a situation through between those sessions or without a coach at all: rehearsing a stop-work conversation before it happens, preparing for a hard one-to-one with a worker who cut a corner, or going back over a debrief that did not go as hoped. As of 2026-10-09, it answers without a booked slot at any hour, which suits shift work, and it draws on relevant detail from your earlier conversations rather than starting blank. It does not observe your site and knows only what you tell it. The concepts it draws on, such as the leader as coach and the after-action review, are published as concept pages you can read before deciding whether any of it is worth your time.

As of 2026-10-09, IX Coach's pricing page lists $40 per month with a 7-day free trial.

Research

Practice this with IX Coach

Try this practice

Keep reading