The Digital HSE Problem Nobody Measures
Most HSE teams now hold more safety data than at any point in their history. Very few have more time to think about it.
Inspection apps, permit-to-work systems and dashboards have multiplied. Yet the HSE professional’s week often looks the way it did ten years ago: chasing overdue actions by email, consolidating spreadsheets before the monthly review and building audit evidence packs the night before the auditor arrives.
The digital HSE debate usually asks whether automation will replace safety professionals. The more urgent question is why so many tools have failed to give those professionals their time, and their judgement, back.
The consequences are not abstract. The ILO estimates that 2.93 million workers died from work-related accidents and diseases in 2019, and that 395 million suffered non-fatal work injuries (ILO, A Call for Safer and Healthier Working Environments, 2023). In Great Britain, HSE UK recorded 138 workers killed in work-related accidents in 2023/24. Behind each figure sits a chain of judgement calls: whether a control was adequate, whether a weak signal was recognised, whether a corrective action worked. A system that does not improve those calls risks becoming little more than a faster filing cabinet.
Key takeaways
- Digital HSE should absorb administration, not professional judgement.
- The Judgement Line splits every HSE task into four zones: Capture, Connect, Challenge and Confirm.
- A practical test of any digital HSE system is whether it increases the share of HSE hours spent on Challenge and Confirm.
Why the Replacement Debate Misreads Safety Work
HSE Software is very good at collecting information, sending reminders, spotting patterns and generating reports. Safety professionals interpret conditions, understand human behaviour and make decisions that depend on context. These capabilities are complementary, not competing.
Consider an incident investigation. Software can organise evidence, hold the timeline, track CAPA owners and retrieve similar events from previous years. It cannot work out why a supervisor signed off a permit with a missing isolation, or whether pressure from a delayed shutdown shaped that decision. That takes interviews, site knowledge and an understanding of how work is actually done.
Risk assessment is the same. A system can supply templates and revision history; it cannot tell whether a JSA written for a day-shift crew still holds at night with a contractor team new to the site.
Site inspections follow the same logic. A digital checklist can confirm that the inspection took place, that photos were attached and that findings were assigned. Whether the inspector noticed that the barricading had not moved with the work front, or that a crew was improvising access to finish before shift change, still depends on the person walking the site.
The regulations already assume this. OSHA 1910.119(m) requires incident investigations to be run by a team that includes at least one person knowledgeable in the process involved. ISO 45001:2018 clause 10.2 requires organisations to review the effectiveness of any corrective action taken. Neither requirement can be fully met by a workflow alone. The opportunity is augmentation, not replacement.
Where the HSE Professional’s Hours Actually Go
Take a common request. An auditor asks for every CAPA raised from contractor incidents in the last 12 months, with closure evidence and proof of verification.
In a spreadsheet environment, three registers get opened: incidents, actions and contractors. The same contractor appears as “ABC Scaffolding”, “ABC Scaff Ltd” and “ABC (Site 2)”. Closure evidence sits in email threads. Verification, where it happened at all, is a free-text comment. Two working days later the pack is ready, and very little of that time involved safety judgement.
The same pattern appears across routine HSE work. In contractor management, prequalification documents, competency cards and insurance certificates are checked by hand against expiry dates, while the more useful question, how each contractor’s crews actually perform on site, gets less attention. In CAPA Management, verification depends on someone remembering to go back, so it is often the step that slips.
For PSM-covered facilities the load is heavier. OSHA 1910.119(o) requires compliance audits at least every three years, with a documented response to each finding. Paragraph (m)(5) requires a system to promptly address and resolve incident findings and recommendations, with resolutions documented.
The cost of displaced attention is real. The NSC estimates the total cost of US work injuries at $176.5 billion in 2023 (NSC Injury Facts). HSE UK puts the annual cost of workplace injury and new cases of work-related ill health in Great Britain at £21.6 billion for 2022/23. Prevention happens at the worksite; excessive clerical work can pull HSE professionals away from the activities where their judgement adds the most value.
The Judgement Line: A Framework for Digital HSE Augmentation
Every HSE task sits on one side of a line. Below it is work defined by rules, repetition and retrieval, which the system should own. Above it is work that needs context, interpretation and challenge, which the professional must own. I call this The Judgement Line.
Digital HSE programmes tend to fall short in one of two ways. Either the system stops short of the line and professionals keep doing clerical work, or it crosses the line, auto-closing CAPAs and generating risk ratings that nobody questions. The Judgement Line has four zones:
Zone | Owner | What it covers | Operational example |
1. Capture | System | Recording hazards, Near Misses, permits, inspections and evidence at the point of work, with time, location and photos attached. | A Near Miss logged from the work front in under two minutes, linked to the live permit. |
2. Connect | System | Linking records, routing actions, escalating overdue items and surfacing repeat events across sites, contractors and assets. | Flagging the fourth dropped-object Near Miss in the same module in 90 days. |
3. Challenge | HSE professional | Testing whether controls fit the real task, whether the RCA reached system causes, and whether each CAPA Management targets the cause or the symptom. | Rejecting “retrain the operator” as the only CAPA for a repeat permit deviation. |
4. Confirm | Shared | The system schedules and evidences verification; the professional decides whether the risk actually fell. | A field check 60 days after closure to see if the new control is in use. |
A simple test places any task: does it require knowledge of the context that is not in the record? If yes, it sits above the line. If no, the system should be doing it.
A practical measure I propose is the Judgement Ratio: the share of HSE team hours spent in Challenge and Confirm compared with Capture and Connect. It is not an established industry KPI, and it does not need to be precise; a two-week time log each quarter is enough to show the direction of travel. If the ratio does not move after a digital rollout, that is worth investigating, because the technology may have digitised the paperwork without returning judgement time to the team.
This is the principle SoapBox is built on: the system carries Capture and Connect so that HSE professionals can spend their hours on Challenge and Confirm.
The Judgement Line in Real Operations
5.1 Repeated work-at-height observations on an EPC site
Fall protection under OSHA 1926.501 has been the most frequently cited OSHA standard for more than a decade, which makes this a familiar pattern on construction and EPC projects. A site inspection records a missing guardrail on a scaffold platform at the pipe rack. It is rectified on the spot and closed. Over the next six weeks similar findings appear: an out-of-date scaffold tag, a harness clipped to a handrail instead of a rated anchor point, a missing toe board at the same elevation. Each is closed as “corrected immediately”.
Individually, every record looks resolved. The system’s job is Connect: it groups observations by activity, area and subcontractor, and shows that seven work-at-height findings in six weeks trace back to one scaffolding subcontractor, most of them on night shift.
The HSE professional’s job is Challenge. On-the-spot correction fixes the condition, not the control. Walking the work front and talking to the crews reveals the underlying gap: scaffolds are being handed over at shift change without inspection by a competent scaffold inspector, and the work-at-height permit does not require the issuer to confirm a tagged scaffold or a verified anchor point. The CAPA moves upstream: scaffold inspection and tagging becomes a hold point in the permit, and the subcontractor’s night-shift supervision is reviewed.
Confirm is shared. The system schedules targeted inspections over the next four weeks and tracks whether work-at-height findings for that subcontractor and area fall. The professional decides whether a drop reflects a working control or simply fewer inspections on nights.
5.2 A SIMOPS Near Miss that looked routine
On an offshore platform, a crane lift runs in parallel with scaffolding work on the deck above. A spanner falls from the scaffold into the lay-down area during the lift. Nobody is hurt. The near miss is captured from a phone with a photo, location and both permit numbers attached. The system connects it to two earlier dropped - object Near Misses during SIMOPS windows in the same area, both closed with a toolbox talk.
Here the HSE professional takes over. Three events with the same closure mean the toolbox talks are treating a symptom; the real gap is a SIMOPS matrix with no exclusion zones for overhead work during lifts. The CAPA becomes a revised matrix, and Confirm means a verification walk during the next three concurrent lifts.
5.3 A temporary MOC that quietly became permanent
A high-level trip on a separator is bypassed under a temporary MOC while an instrument is repaired. The system routes approvals, records a manual level watch as the compensating measure and sets an expiry date. The judgement call is whether a manual watch is realistic across a 12-hour night shift.
Weeks later the spare part is still on back order and the MOC passes its expiry. The system escalates it automatically. Under API RP 754, an overfill with loss of primary containment could become a Tier 1 or Tier 2 PSE depending on the quantity released. Overdue temporary MOCs, tracked as a Tier 4 operating discipline indicator, show that drift long before it reaches the release.
5.4 The contractor CAPA audit, revisited
With Capture and Connect handled by the system, the auditor’s request takes minutes. The professional’s time moves to the auditor’s next question: how many of those CAPAs were verified effective, and how many were simply marked closed?
Reporting Culture: The Test Every Digital HSE System Must Pass
If workers believe technology is used for surveillance or punishment, Near Miss reporting falls. If they see it as a tool for finding and fixing hazards, participation rises. ISO 45001 clause 5.4 requires organisations to determine and remove obstacles or barriers to worker participation, and the standard’s own notes list reprisals or threats of reprisals among them. A few design choices matter more than any feature list:
- Anonymous route: offer one for Near Miss Reporting and unsafe conditions.
- Closed loop: every reporter should see what was done with their report.
- Separation: keep Near Miss data apart from disciplinary processes by default.
- Culture indicator: track the time from a Near Miss report to feedback to the reporter.
When workers stop reporting because they fear how the data will be used, the dashboard may look healthier while the underlying risk becomes less visible.
The Two Skill Sets of the Future HSE Professional
The HSE professional of the next decade needs two skill sets. The first is unchanged: hazard recognition, RCA methods, human factors, and a working command of PSM elements such as those set out in the CCPS Risk Based Process Safety framework. The second is digital literacy: understanding data structures and taxonomies, reading dashboards critically, designing workflows, and scrutinising AI-generated outputs. In practice, that means being able to tell whether an inspection checklist is capturing the hazards that matter on a given site, or only the ones that are easy to tick.
When an AI tool proposes a root cause or a risk rating, treat it as a hypothesis to Challenge, not a conclusion to accept. The goal is not to make safety more technological for its own sake. It is to make safety management more informed, timely and effective.
The Question Excel Cannot Answer
The best digital HSE system will not replace the person who walks onto the worksite and recognises a serious hazard. It will help that person act faster, document better, identify patterns earlier and demonstrate whether the control actually worked. An EHS Software systems can support that work by connecting the informtion needed to make those decisions.
That last part is where spreadsheets run out. Excel can tell you how many CAPAs you closed last year. It cannot, on its own, tell you this:
“Of the corrective actions we closed in the last 12 months, which ones actually reduced the risk they were raised to control, and what evidence proves it?”
Answering it requires an unbroken chain from Near Miss to RCA, from RCA to CAPA, from CAPA to verification, and from verification to recurrence data. That chain is exactly where the Judgement Line sits, and it is a question every HSE leader should be able to answer before the next audit asks it.
About the author
Hasan Ahmed Sayed is an HSE professional with 12+ years of experience across construction, EPC, industrial, shutdown and turnaround projects. Currently working as an HSE SME at SoapBox Cloud Software Solutions LLP, he combines practical HSE field experience with technology to help translate real-world safety requirements into effective digital HSE solutions and workflows.
References
1. ILO (2023). A Call for Safer and Healthier Working Environments. https://www.ilo.org/publications/call-safer-and-healthier-working-environments
2. HSE UK. Work-related fatal injuries in Great Britain. https://www.hse.gov.uk/statistics/fatals.htm
3. HSE UK. Costs to Britain of workplace injuries and new cases of work-related ill health. https://www.hse.gov.uk/statistics/cost.htm
4. NSC Injury Facts. Work injury costs. https://injuryfacts.nsc.org/work/costs/work-injury-costs/
5. OSHA. 29 CFR 1910.119, Process Safety Management of Highly Hazardous Chemicals. https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.119
6. OSHA. 29 CFR 1926.501, Duty to have fall protection; and Top 10 Most Frequently Cited Standards. https://www.osha.gov/top10citedstandards
7. ISO 45001:2018. Occupational health and safety management systems. https://www.iso.org/standard/63787.html
8. API RP 754 (3rd ed., 2021). Process Safety Performance Indicators for the Refining and Petrochemical Industries. https://www.api.org/
9. CCPS (2007). Guidelines for Risk Based Process Safety. https://www.aiche.org/ccps
