Safety culture is tested in the moment a schedule slips, an alarm sounds, a contractor challenges a control, or a senior manager asks for an accelerated handover. Knowing how to strengthen safety culture means addressing what people actually do when competing pressures are real – not restating the standards they can already recite.
In high-risk and highly regulated environments, safety is not sustained by policy volume or annual messaging. It is sustained by the quality of operational judgement, the consistency of leadership behaviour, and the organisation’s ability to spot drift before it becomes an event. That requires a culture programme built around observable work, not aspiration.
Start with the behaviours that shape risk
Most organisations say safety is a value. The more useful question is whether that value can be seen in decisions, conversations and handovers. If it cannot, leaders cannot reinforce it reliably and teams cannot hold each other to account.
Define the few behaviours that matter most in your operating context. These may include stopping work when conditions change, escalating uncertainty early, challenging a weak control, conducting a proper peer check, or making time for an effective pre-job brief. The right set will differ between a nuclear site, a defence programme and a major engineering project. The principle does not: make safe performance specific enough to observe.
Avoid turning this into another competency framework with dozens of statements. Under pressure, people need practical cues. What does a good challenge sound like? When is a pause mandatory rather than optional? What information must be shared before a decision can proceed? Clear behavioural standards reduce the space in which risk can be rationalised as speed, experience or commercial necessity.
How to strengthen safety culture through leadership behaviour
Leaders create the conditions in which people decide whether speaking up is worthwhile. Their influence is strongest not in formal safety briefings, but in the small operational moments that teams notice: how they react to bad news, whether they ask about uncertainty, and whether they respect a decision to stop.
A leader who says “safety first” but becomes visibly impatient when a team raises a concern teaches a different lesson. So does a project director who accepts incomplete assurance to protect a milestone. Teams are highly capable of reading the real priority, particularly when workload and delivery pressure increase.
Leadership visibility therefore needs to be more than site presence. Senior people should spend time where work is planned, controlled and handed over. Their questions should test the strength of controls and decision-making, rather than invite rehearsed reassurance. Ask what has changed since the work was authorised, where people are relying on assumptions, and what would make the task unsafe to continue.
This is not about leaders becoming technical specialists in every activity. It is about demonstrating disciplined curiosity and making it legitimate for others to do the same. In complex operations, credible leadership means knowing when to enquire, when to challenge and when to hold the line.
Make challenge a normal operating practice
Psychological safety is often discussed too broadly to be useful. In safety-critical work, the requirement is more precise: people must be able to raise a concern, challenge a decision and report a near miss without being dismissed, isolated or quietly penalised.
That does not mean every challenge is correct or every decision is resolved by consensus. Leaders still need to make decisions, and operations still need pace. But people need evidence that a challenge will receive a proportionate response. Acknowledgement, an explanation of the decision, and visible follow-through are usually more valuable than a generic invitation to speak up.
Pay particular attention to junior staff, temporary workers and contractors. They often see issues first while having the least organisational confidence to raise them. If their route to escalation is unclear, slow or politically risky, the organisation loses valuable early warning.
Treat reporting as intelligence, not administration
A weak reporting culture is rarely fixed by asking for more reports. People report when they believe the information will be used intelligently and fairly. If reports disappear into a system, result in blame, or create paperwork with no visible outcome, reporting will become selective.
Near misses, workarounds, repeated minor deviations and concerns raised in briefings are signals of how work is really being done. They can reveal overloaded controls, unclear interfaces, poor planning or a growing gap between procedure and practice. The objective is not simply to increase the number of entries. It is to understand patterns early enough to intervene.
Close the loop visibly. When someone raises a concern, explain what was found, what changed and why. Where no action is taken, explain that too. This builds trust in the process and improves the quality of future information.
Fair accountability matters here. Deliberate disregard for controls, reckless conduct and repeated non-compliance require firm action. Human error, however, should prompt examination of conditions, system design, supervision and workload. Treating every mistake as an individual failure drives risk underground. Treating every failure as a system issue can remove personal responsibility. Effective safety culture holds both truths at once.
Build safety into planning, not recovery
Safety degradation often begins well before the point of work. It begins when scopes are rushed, interfaces are poorly defined, resources are stretched or changes are accepted without adequate review. By the time a supervisor is managing the consequences on the ground, the available choices may already be constrained.
Senior teams should examine where operational pressure is generated. Are programme milestones producing unrealistic sequencing? Are teams working around recurring permit or access problems? Are changes being approved with incomplete information? Are contractor incentives aligned with safe delivery, or do they reward apparent progress regardless of rework and risk?
This is where safety and performance should not be treated as opposing objectives. Poorly planned work is not efficient work. It creates rework, delay, fatigue, conflict and increased exposure to error. Reliable delivery depends on leaders recognising that controls, time for preparation and quality of handover are performance requirements.
Measure the leading indicators of cultural drift
Lagging data matters, but it tells leaders what has already happened. Incident rates, lost-time injuries and regulatory findings should never be ignored, yet they provide limited visibility of weakening conditions before an event.
Use leading indicators that reflect the behaviours and conditions you are trying to improve. This may include the quality of pre-job briefs, the timeliness of escalation, recurring control failures, unresolved concerns, leadership field engagement, decision turnaround at critical interfaces, or the extent to which teams feel able to challenge. The measure must be credible to the people doing the work. If it is easily gamed, it will become a reporting exercise rather than an improvement tool.
Data also needs context. A rise in reported near misses may indicate growing risk, but it may equally show that people trust the reporting process more than they did previously. Leaders should avoid simplistic interpretations and look for patterns across locations, teams, shifts and contractors. Quantitative data is stronger when tested against direct observation and structured conversations.
TEG’s approach is built on this closed loop: develop the leadership behaviours required in live operating conditions, then use behavioural insight to see whether those behaviours are holding. Without that feedback, organisations can confuse attendance at development activity with cultural change.
Reinforce expectations in the operational rhythm
Culture is not changed in a launch event. It is reinforced through the routine cadence of work: planning meetings, shift handovers, operational reviews, project boards and leader walkabouts. Each forum should make room for risk, uncertainty and learning in a way that is disciplined rather than ceremonial.
For example, a project review should not only ask whether delivery is on track. It should ask which assumptions have changed, where controls are fragile, what concerns remain unresolved and who owns the next decision. A handover should not simply transfer task status. It should transfer the risks, constraints and judgement behind the status.
Consistency across levels is critical. A frontline supervisor cannot sustain a high standard of challenge if middle managers reward only output, or if executives override local decisions without understanding the controls. Safety culture strengthens when people encounter the same expectations from boardroom to workface.
The practical test is straightforward. When pressure rises, does the organisation become quieter, faster and more defensive? Or does it become clearer about risk, more willing to challenge assumptions and more disciplined in its decisions? Build for the second response, because that is where safety culture proves its value.