How a Global Safety Transformation Portfolio Replaced Audit Completion With Decision Evidence

The audit report was complete, the action log was green, and the leadership team was ready to close the review. Then a supervisor asked a question that changed the conversation. "Which decision is different in the field because we completed this audit?"

The question exposed a familiar gap. The organization had evidence that inspections, meetings, training sessions, and corrective actions had been recorded. It had less evidence that a serious exposure had become harder to encounter, easier to detect, or more likely to trigger a timely decision.

Across more than 250 cultural transformation projects, Andreza Araujo's work points to a practical distinction: an audit becomes valuable when its findings change ownership, resources, work conditions, or verification. The transformation is not from paper to digital reporting. It is from completed activity to visible decision evidence.

This case study describes a recurring transformation pattern rather than inventing a result for one unnamed plant. Its scope is grounded in Andreza Araujo's verified experience serving more than 250 companies worldwide, with impact in more than 30 countries, and in her executive work across large multinational operations. The examples show what changed in the management system and how leaders can test whether the change is real.

Initial scenario: the organization measured completion better than control

The starting point was not a careless operation. Most organizations in this situation had standards, trained professionals, reporting channels, and leaders who genuinely wanted better safety performance. Their weakness appeared in the distance between the activity recorded and the decision required.

An audit could show that a permit had been reviewed, a supervisor had completed a conversation, or a corrective action had been marked closed. Those facts mattered, although they did not answer whether the control held when production pressure increased, a contractor changed the work method, or the field condition no longer matched the original plan.

OSHA's Recommended Practices for Safety and Health Programs emphasizes management leadership and worker participation, while the HSE's HSG65 guidance connects planning, doing, checking, and acting. Both references support a management cycle in which information must lead to action. A completed record is only one part of that cycle.

James Reason's work on organizational accidents helps explain why the gap persists. A visible event can be preceded by latent conditions in design, supervision, maintenance, information flow, or resource allocation. When a review examines only whether a procedure exists, it can miss the conditions that make the procedure unreliable.

The decision: make evidence answer an operational question

The transformation began when leaders stopped asking whether the system was active and started asking what decision the evidence should support. "Are we compliant?" was replaced with questions such as, "Can this control prevent the exposure during this task?" and "What should the supervisor do when the control is unavailable?"

That change gave every review a sharper boundary. The team named the exposure, described the control that should prevent or limit it, identified the owner who could change the condition, and agreed on the evidence that would show whether the action worked.

Andreza Araujo's book Safety Culture: From Theory to Practice treats culture as a pattern of management choices that people can observe. The same principle shaped this transformation. A value statement could remain in the policy, but the review moved toward decisions that workers could see in equipment, scheduling, supervision, authorization, and escalation.

The decision also protected EHS from becoming the sole owner of improvement. EHS could define the method, challenge weak evidence, and help leaders interpret patterns. Operations still had to own the exposure and provide the resources required to change it.

Execution: convert diagnosis into a field decision loop

The execution pattern used a repeatable loop with enough structure to compare sites and enough flexibility to respect local conditions. It started with diagnosis, but diagnosis was not treated as the deliverable.

First, leaders selected a small number of exposures that could create serious harm and wrote the decision associated with each one. A working-at-height review, for example, had to show more than training completion. It had to show whether anchor points, rescue arrangements, supervision, and authorization were usable for the work planned that day.

Second, the team separated reported activity from field evidence. A training record could remain useful, but it was paired with observation of the decision the training was meant to support. An audit closure could remain in the system, but closure also required evidence that the work condition had changed and that the change survived a later verification.

Third, leaders assigned the response to the person with authority over the condition. When the issue involved equipment design, the owner was not asked to solve it with another briefing. When the issue involved a production constraint, the corrective action included the operational decision needed to remove or manage that constraint.

Finally, the review returned to the field. The question was not whether the team had followed the improvement process. The question was whether a worker, supervisor, or manager would now make a different choice when the exposure appeared.

Measured result: the portfolio created a wider evidence base

The measurable scope of this case is the portfolio itself. Andreza Araujo has led or supported more than 250 cultural transformation projects across more than 30 countries, after more than 25 years in multinational EHS and safety culture roles. That breadth does not justify assigning one identical result to every project. It does show that the evidence question can be applied across industries, operating models, and levels of maturity.

A second verified reference point came from Andreza's PepsiCo South America experience, where an 180-day plan led to a 50% reduction in accident ratio in six months. That result is not presented as an average for the 250-plus projects. It is a separate example of why leadership routines, operational ownership, and field execution matter more than a dashboard that only reports completion.

The transformation therefore produced two kinds of measurable value. The first was operational scope, with the method used across a substantial international portfolio. The second was decision quality, shown when a review could trace a finding to a changed condition, a named owner, and a later effectiveness check.

What changed in leadership behavior?

The first change was the quality of the question. Leaders became less satisfied with "How many actions are overdue?" and more interested in "Which exposure remains possible, and who has the authority to change it?" The second question is harder, but it is closer to the risk.

The second change was the tolerance for weak closure. A green status no longer ended the conversation when the supporting evidence was a signature, a photograph without context, or a repeated training session. Leaders asked what the control looked like in the work and what would reveal a regression.

The third change was the treatment of bad news. A system that rewards only clean reports will eventually receive clean reports instead of accurate information. Worker participation and protection from retaliation, emphasized in OSHA's guidance, become practical requirements when leaders need evidence from the field.

The fourth change was cadence. Evidence was reviewed close enough to the work that a weak control could be corrected before it became normalized. That rhythm mattered because culture is not created by one campaign. It is reinforced by what leaders repeatedly approve, challenge, fund, and verify.

Generalizable lessons for safety culture leaders

Four lessons transfer well across sectors. First, diagnose the operating system, not only the opinions around it. Surveys and interviews can reveal perception, but the decisive evidence often sits in work allocation, equipment condition, authorization, and response to exceptions.

Second, define the evidence before assigning the action. If the team cannot state what will demonstrate effectiveness, the action is still an assumption. A useful acceptance criterion describes the changed condition, the person who will verify it, and the date or trigger for verification.

Third, keep the result proportional to the claim. If a program served 250-plus companies, say that. If one regional operation reduced accident ratio by 50% in six months, name that specific case. Do not turn a verified example into a universal promise.

Fourth, make the next decision visible to workers. If the transformation changes only the dashboard, the field may never experience it. When the change affects a permit, a staffing decision, an equipment purchase, a maintenance priority, or a stop-work response, the culture shift becomes observable.

What to apply in your operation

Select one high-consequence exposure and review the last five actions associated with it. For each action, ask whether the record proves activity, a changed condition, or effective control. Then remove one measure that does not change a decision and replace it with evidence that an operational owner can act on.

Use a simple four-part record. Name the exposure, identify the critical control, assign the decision owner, and define the verification evidence. When the evidence is missing, the action is not ready to close. When the control is unavailable, the leader must decide whether to correct, redesign, postpone, or stop the work.

Andreza Araujo's approach in Safety Culture Diagnosis: Learn how to do your own reinforces the value of diagnosis that leads to action. A diagnosis should make the next decision clearer, not merely produce a more detailed report.

For leaders building this discipline across sites, Andreza Araujo's safety culture and leadership work offers a deeper brand-aligned reference. The practical test remains simple: can the organization show what changed in the work, who made the decision, and how it knows the change lasted?

Frequently asked questions

A safety transformation becomes credible when evidence from the field changes a decision, an owner, a resource, or a work condition, rather than only increasing the number of completed activities.

What is decision evidence in safety culture?

Decision evidence is information that shows whether a safety-related choice changed because of what the organization learned. It can include a redesigned control, a changed authorization rule, a resourced corrective action, or a verified decision to stop or postpone work.

Does a completed audit prove that risk is controlled?

No. An audit can identify a gap and confirm that a review occurred, but risk control still requires evidence that the relevant condition changed and that the control works under the operating conditions where people face the exposure.

Who should own a safety culture transformation?

Senior leaders should set direction and provide resources, while operational leaders should own the conditions that create or reduce exposure. EHS should provide technical method, assurance, and challenge without becoming the only function responsible for changing the work.

How can a small site use this approach?

Start with one serious exposure, one accountable operational owner, one field verification, and one leadership review. The method can remain compact when the decision, evidence, and escalation route are explicit.

What is the most important first question?

Ask, "Which decision is different because of this safety activity?" If the answer is unclear, the activity may still be useful, but it should not be reported as proof that the exposure is controlled.