How Cam Stevens Thinks About Problem-First Technology

Episode 15 of Headline Podcast featured Cam Stevens, safety technologist and CEO of PKG, in a conversation published on March 18, 2026. Stevens defended a practical thesis for EHS leaders: technology becomes useful only after the organization defines the risk problem it expects the tool to change.

Problem-first technology is the discipline of defining the safety decision, field evidence, worker trust condition and control owner before selecting any digital tool that could create alarm dependency. In EHS, it prevents leaders from buying software, sensors or AI systems that increase reporting volume without improving risk control.

Why does Cam Stevens start with the problem?

Cam Stevens starts with the problem because a technology catalog cannot decide which risk matters, which field signal deserves attention or which leader must act within 24 hours. In Episode 15, he says the danger begins when a team shops for a tool before naming the operational decision it wants to improve.

His clearest line on the show was direct: "If we're very clear on the problem to solve, then a technology catalog is excellent. If we're not clear on the problem to solve, a technology catalog is very dangerous." That sentence matters because many EHS rollouts begin with vendor demos, not with the exposure pathway, decision delay or control weakness that the organization must fix.

OSHA explains that effective safety and health programs use a proactive approach to find and fix hazards before harm occurs. Stevens pushes that idea into technology adoption: the first test is not whether the product is impressive, but whether it helps the company find and fix a specific hazard faster than the current system.

Across 25+ years leading EHS at multinationals, Andreza Araujo has seen the same pattern in technology rollouts: a dashboard can create speed without judgment, and speed without judgment only moves weak decisions through the system faster.

What changes when the tool is selected before the risk decision?

When the tool comes before the risk decision, the organization usually measures adoption rather than control. A 90 percent app usage rate can still hide the fact that supervisors are not closing high-risk actions, contractors are not escalating changes, and senior leaders are not seeing the signals that should alter work.

This is why technology catalogs can turn safety tools into risk when leaders confuse procurement with governance. The hidden failure is not the software. The hidden failure is a missing sentence that should have existed before purchase: this tool must improve this decision, by this owner, inside this response window.

4 controls should be named before selection: the risk decision, the evidence source, the owner of response and the verification method. If one of those 4 is missing, the technology can still look modern while leaving the control chain unchanged.

As Andreza Araujo argues in Safety Culture: From Theory to Practice, culture appears in repeated decisions, not in slogans. A digital system that does not change repeated decisions is not culture change. It is a new interface over the old operating habit.

How should EHS leaders define the problem before a pilot?

EHS leaders should define the problem before a pilot by writing one decision statement, one field evidence statement and one action statement. For example, a 300-employee plant might define the problem as delayed recognition of maintenance risk during temporary work, then test whether the tool shortens the time from worker signal to supervisor action.

The pilot brief should avoid vague goals such as improving engagement or modernizing safety. It should say what current decision is late, what evidence is missing, whose judgment is weak and what consequence follows when nothing changes. Without that discipline, adoption data becomes the easiest number to report because it says nothing about risk.

NIOSH describes the hierarchy of controls as a way to protect workers by ranking control methods from more effective to less effective. That hierarchy is a useful discipline for safety technology because the tool should support control selection and verification, not distract the team with data volume.

The practical move is simple enough for a 30-minute governance meeting. Ask the sponsor to complete 3 sentences before any pilot begins: the risk we are trying to reduce is specific, the decision that currently fails is named, and the field proof that will show improvement is observable.

Why is prediction hype weaker than field signal quality?

Prediction hype is weaker than field signal quality because a model cannot overcome poor inputs, weak trust or delayed leadership response. In a safety setting, the useful question is not whether technology can predict something, but whether the organization can act on a credible signal while the control window is still open.

Stevens' episode connects directly to voice technology as safety signal governance. Voice tools can reveal weak signals from the field, but only when workers believe the data will be used to fix conditions rather than monitor blame, productivity or loyalty.

BLS records fatal and nonfatal occupational injury and illness data through its Injuries, Illnesses, and Fatalities program, but national lagging data cannot tell a supervisor what to do at 9:10 a.m. during a changing task. A good technology pilot must connect both levels: strategic pattern recognition and immediate field action.

During the PepsiCo South America tenure, where the accident ratio fell 50 percent in 6 months, Andreza Araujo learned that the decisive shift was not a prettier report. The decisive shift was faster ownership of the signal, with leaders treating weak evidence as a prompt for field verification rather than as a debate to postpone.

What did Cam Stevens say about psychosocial risk?

Cam Stevens warned that technology will reshape psychosocial exposure, not only physical risk. His Episode 15 statement was blunt: "The changing shift in risk profile will be overwhelmingly psychosocial, driven by technology usage in our organizations." That matters for EHS because digital systems can change workload, monitoring pressure, autonomy and trust.

This is the trap most technology business cases minimize. A tool can reduce one operational friction while increasing interruption, surveillance anxiety or decision fatigue. The issue is not anti-technology sentiment. The issue is whether leaders examine the work design consequences before workers absorb them.

That is why psychosocial risk from technology needs a pre-rollout audit, especially when the system changes how work is assigned, timed, scored or escalated. ISO 45003 gives leaders a language for psychosocial risk, but the field test is concrete: does the tool make work clearer, or does it add another demand without removing an old one?

30 days is enough for a first psychosocial review when the pilot team interviews workers, supervisors and HR about workload, clarity, monitoring pressure and recovery. If the review waits 6 months, the new stressor may already feel normal.

How does problem-first technology protect worker trust?

Problem-first technology protects worker trust by telling people why data is collected, what decision it supports and what will not be done with it. In Episode 15, Stevens added a boundary that every EHS leader should keep visible: "We can elevate the human experience with technology, but there are certainly times when technology should absolutely not be used."

Trust is not a communication accessory after rollout. It is part of the control design. If workers believe a voice tool, wearable or AI triage system will be used to punish normal reporting, the signal source collapses, and the organization receives cleaner data precisely because people stopped saying risky things.

On the Headline Podcast, co-hosted with Dr. Megan Tranter, Andreza explored this with Stevens as a governance problem, not only as a product problem. The leader has to define non-use boundaries, privacy expectations, escalation rules and appeal paths before the first metric appears in a dashboard.

The field question for a pilot team is uncomfortable but necessary: would a worker still report a weak signal if the supervisor, plant manager and corporate team could all see the data? If the answer is no, the tool is not ready for the trust conditions it needs to work.

Problem-first technology versus technology-first rollout

The difference between problem-first technology and a technology-first rollout is visible before purchase. Problem-first teams define the risk decision, affected work, response owner and verification evidence first, while technology-first teams define features, integrations, licenses and adoption targets first. Both can produce activity, but only one starts from control.

ISO specifies ISO 45001:2018 as an occupational health and safety management systems standard, which means technology should support a management system rather than become a separate novelty layer. The same logic applies to AI in EHS: the decision process must govern the tool, not the other way around.

This distinction also explains why AI in EHS requires executive ownership. If executives delegate the adoption decision entirely to IT or procurement, they also delegate the ethical, operational and cultural consequences that appear when the system meets real work.

Recommendation

EHS leaders should approve safety technology only after the pilot charter names 5 things: the risk problem, the decision owner, the field evidence, the worker trust boundary and the verification cadence. If the sponsor cannot write those 5 elements in plain language, the organization is not ready to buy the tool.

The strongest practical sequence is to run a 2-week problem definition sprint before vendor selection. Interview supervisors, workers, maintenance, HR and operations. Identify where the current signal fails, then test whether technology can improve that failure without creating a new psychosocial burden.

Each month spent chasing a tool before defining the problem creates a quieter form of risk: leaders become busy with implementation while the original exposure remains unmanaged in the field.

Andreza Araujo's work across more than 250 cultural transformation projects points to the same conclusion. Technology can accelerate a mature decision system, but it rarely rescues a weak one. The first investment is not the platform. The first investment is decision clarity.

Conclusion

Cam Stevens' Episode 15 argument is not anti-technology. It is pro-discipline, because the safest digital rollout starts with the risk decision that must improve, the people whose trust makes the signal possible and the leaders who must act when the signal arrives.

Listen to the full conversation: Listen to the full conversation.