New Safety Dialogue Facilitator in 60 Days: Build Conversations That Change Work

F6 role profile for a new safety dialogue facilitator

A safety conversation can last fifteen minutes and still leave the work exactly as it was. The team may describe a shortcut, an unreliable guard, a rushed handover, or a conflict between production and protection, while the facilitator records the exchange as a completed observation.

That is the central risk for anyone taking responsibility for safety dialogue. The role can become a scheduling function that counts conversations, distributes cards, and reports participation. It becomes useful only when the conversation changes a decision, a control, a work sequence, or the route by which an unresolved concern reaches authority.

A safety dialogue facilitator helps people examine the conditions that shape a safety decision, convert the evidence into an owned response, and verify whether the response changed the exposure. The facilitator improves the quality of the conversation without replacing the supervisor or manager who owns the work.

James Reason’s work on latent conditions and active failures remains important here because the visible action is only one part of the explanation. A person may accept a risky sequence, yet the decision can be shaped by access, equipment design, staffing, time pressure, unclear priorities, or a supervisor response that has rewarded the same shortcut for months.

Andreza Araujo’s approach to safety culture treats this gap between the written expectation and the repeated decision as a leadership problem. Across more than 25 years in multinational EHS, her work has consistently connected culture with what leaders permit, reinforce, measure, and change. The facilitator’s first 60 days should make that operating reality visible.

What does a safety dialogue facilitator need to understand before starting?

The facilitator is not the owner of every problem discovered in conversation. The role is to make the evidence precise, help the right person decide, and protect the follow-through until the exposure has been reviewed.

Begin by separating four things that are often mixed together. The first is the task people are expected to perform. The second is the task as the worksite actually allows it to happen. The third is the exposure created by the difference. The fourth is the decision required to close the difference.

A dialogue that asks only, “Did you follow the procedure?” usually produces a predictable answer. A stronger conversation asks which part of the procedure was difficult to apply, what made the alternative attractive, which control was available, and what would have to change before the same decision appears again.

ISO 45001:2018 gives consultation and participation a management-system place, but the standard does not prescribe a script that makes people speak honestly. ISO 45003:2021 adds useful attention to work organization and psychosocial conditions, which matters when the quality of a conversation is affected by fear, fatigue, conflict, or low control.

Write a one-sentence role boundary before the first field visit. A practical version is, “I help the team describe the exposure and route the decision; the accountable leader owns the change and its verification.” That sentence prevents the facilitator from becoming either a passive note-taker or an unofficial owner of every corrective action.

First 14 days: map where conversations become ritual

The first two weeks should produce a map of the current dialogue system, not a new campaign. Trace what happens before, during, and after one conversation.

Review the forms, prompts, training materials, observation cards, supervisor instructions, and escalation records already in use. Do not judge the system by how polished the documents look. Select a small sample of completed conversations and ask whether a person who was not present could understand the exposure, the decision, the owner, and the later result.

Shadow conversations across different conditions, including a normal shift, a handover, a maintenance interruption, and a period of production pressure. The point is not to compare personalities. It is to see whether the method still works when the team has less time, more uncertainty, or a competing business priority.

Use three diagnostic questions with supervisors. What do you do when a worker identifies a problem you cannot fix today? Which concerns return repeatedly? What evidence tells you that a conversation changed the task rather than simply improved the record?

Record the answers as system observations. A recurring concern with no owner is a decision-rights problem. A concern that is closed without field verification is an evidence problem. A concern that workers stop raising is a trust problem, even if the dashboard appears cleaner.

Days 15 to 30: build a conversation that can carry evidence

A useful dialogue moves from the task, to the exposure, to the decision, and then to the evidence that will show whether the decision worked.

Teach facilitators to begin with the work, not with the person. Ask the worker to describe the sequence, the point at which the task becomes difficult, the control that should be available, and the condition that makes the workaround seem reasonable. This order reduces defensiveness because it examines the operating context before evaluating personal conduct.

Then test the control. Is it present? Is it accessible? Does it work under the current condition? Does the worker understand its purpose? Can the supervisor make it available without creating a different exposure? These questions are more useful than a generic request to “be more careful,” because they force the organization to distinguish a behavior choice from a design or management weakness.

Charles Duhigg’s work on habits helps explain why a repeated shortcut can persist even after training. A cue, a routine, and a reward can stabilize a behavior that the procedure rejects. In a safety dialogue, the facilitator should look for the reward structure, which may be a faster restart, approval from a senior operator, or relief from an impossible queue.

End every conversation with a disposition. The issue may need immediate containment, technical review, supervisor action, a management decision, or no further action because the evidence does not support the concern. Each disposition needs an owner and a review date. Without that final decision, the dialogue has created disclosure without response.

Days 31 to 45: coach supervisors to respond without closing the conversation

Workers learn whether speaking up is worthwhile from the first response they receive, not from the organization’s policy statement.

Coach supervisors to acknowledge the concern before defending the current method. Acknowledgement does not mean agreement, and it does not remove accountability. It means the person can see that the information reached a decision-maker and will not be dismissed because it arrived at an inconvenient time.

Give supervisors a short response sequence. State what was heard. Clarify the exposure. Say what can be changed now. Name what requires escalation. Explain when the person will receive an update. The sequence works because it replaces vague reassurance with a visible process.

The facilitator should also teach the boundary between learning and tolerance of conscious transgression. A worker who raises a conflict between the procedure and the physical setup deserves investigation of the conflict. A person who deliberately bypasses a known critical safeguard still requires a clear accountability response. Psychological safety makes the concern speakable; it does not make every action acceptable.

Review two conversations per supervisor during this phase. Look for questions that explore conditions, statements that identify authority, and evidence that the response was communicated. Do not score warmth as the main outcome. Score whether the response improved the next decision.

Days 46 to 60: connect dialogue to control verification

The role becomes credible when leaders can see which conversations changed the work and which ones exposed a barrier that still needs a decision.

Build a simple evidence trail with five fields. Capture the task or condition, the exposure, the immediate response, the accountable owner, and the verification result. These fields are enough to stop the process from collapsing into a count of conversations while still keeping the routine practical for supervisors.

Verify a sample of closed items in the field. Speak with the people who perform the task, inspect the control, review the relevant record, and ask what happens when the control is unavailable. A closed action that exists only in a spreadsheet is not yet evidence of control.

Compare repeat concerns with response time. A repeated concern may indicate that the technical action was too weak, that the decision owner lacked authority, or that the response was never communicated to the next shift. Do not assume that a lower count means improvement, because silence can also reduce the number of reports.

Use the results in a short monthly review with operations, EHS, and the relevant technical owners. The review should ask which barrier needs a decision now, which action needs stronger verification, and which supervisor routine should change. That cadence makes dialogue part of control management rather than a parallel engagement program.

Common mistakes that weaken the role

Most weak dialogue programs fail because they confuse participation with influence, or because they ask people to speak without improving the response system.

The first mistake is measuring volume as success. More conversations can mean better access to workers, but it can also mean that the organization is collecting more observations than it can resolve. Participation becomes meaningful when the resulting evidence reaches an owner and returns to the people who raised it.

The second mistake is making the facilitator responsible for fixing every exposure. That arrangement hides management failure and creates a bottleneck. The facilitator should protect the decision path while the operational owner controls the change.

The third mistake is treating every disagreement as a behavior problem. James Reason’s latent-condition lens is a useful safeguard because it requires the organization to inspect the system around the action. The facilitator should ask whether the equipment, layout, staffing, sequence, incentive, or supervision made the expected behavior difficult.

The fourth mistake is using a script so rigidly that people stop describing what is real. A consistent structure is valuable, but the facilitator should follow the evidence when the conversation reveals a design flaw, a conflict of authority, or a serious concern outside the original prompt.

Resources to deepen the role

The facilitator should study both the technical management system and the human conditions that shape a decision at the point of work.

Start with ISO 45001:2018 for consultation, participation, operational control, performance evaluation, and improvement. Use ISO 45003:2021 when fatigue, workload, conflict, or fear affects whether people can raise and resolve risk. Review James Reason’s Human Error for latent conditions and organizational defenses, and use Charles Duhigg’s The Power of Habit to examine why shortcuts become stable routines.

Andreza Araujo’s Safety Culture: From Theory to Practice helps connect the conversation with the repeated decisions that reveal culture. Make The Difference: Be a Leader in Health & Safety supports the supervisor’s role in turning concern into action, while Vamos a Hablar? adds a practical focus on dialogue that makes risk visible without turning the exchange into a ritual.

Safety School courses can deepen the facilitator’s practice, but formal learning should be paired with field review. The role is ready to mature when the facilitator can show a complete chain from conversation to decision, from decision to changed work, and from changed work to verified evidence.

If your organization needs to turn safety dialogue into visible operating practice, Andreza Araujo helps leaders connect safety culture, leadership, and measurable prevention.

What success looks like after 60 days

After 60 days, success is not a larger conversation count. It is a more trustworthy decision path.

The facilitator should be able to show where dialogue happens, which questions reveal the work as performed, how concerns are assigned, and what happens when the owner cannot act immediately. The team should know how to raise a conflict, the supervisor should know how to respond, and leaders should see which barriers require a decision beyond the worksite.

A strong 60-day review includes the open items that still need authority, the closed items that passed field verification, the repeat concerns that deserve a deeper technical response, and the conversations that changed a routine before harm occurred. It also states what the evidence cannot yet prove.

Safety dialogue is not a softer version of control. It is one way to discover whether a control can survive the conditions in which people actually work. When the facilitator protects that link, every conversation has the potential to improve the next decision rather than merely document the last one.