What Does a Veiligheidskundige Do? The Role Explained

The job behind the job title

Ask ten people what a veiligheidskundige does and you get ten answers.

Inspections. Toolbox talks. Incident reports. The RI&E. The folder that gets opened when an inspector walks in. All of those are true, and all of them are incomplete, because they describe tasks rather than the thing the tasks are for.

The core of the role is narrower and harder than the task list suggests: keeping the connection intact between the risks an organisation actually runs, the measures it claims control those risks, and the evidence that those measures still work today. Procedures, rounds, audits and investigations are all instruments serving that one purpose.

Which is why the job is not writing safety down, but keeping what is written and what happens from drifting apart.

The five parts of the work

In almost every organisation, whatever the sector, the role breaks down into five recurring blocks.

• Identifying and assessing risk. RI&E, task risk analysis, job safety analysis, bowties linked to installation scenarios, and the assessment of a change before it goes live rather than after.
• Advising management. Translating a technical finding into a decision someone with budget and authority can actually take, including the uncomfortable version of the advice.
• Designing the system. Procedures, work instructions, permits, training, competence requirements, emergency plans, and the question of who owns each of them.
• Checking. Workplace rounds, asset and installation inspections, internal audits, measurements, and reviewing whether a control measure is still doing what it was designed to do.
• Learning. Near miss and incident reporting, investigation, corrective and preventive actions, trend analysis, and feeding all of it back into the management review.

Plan, do, check, act, in other words. The first two blocks get most of the attention and most of the budget. The last two are where the role is usually judged, and where it usually hurts.

MVK and HVK: what the levels mean

Dutch practice distinguishes between the MVK, middelbaar veiligheidskundige, and the HVK, hoger veiligheidskundige. The rough split: an MVK works close to the operation, on the floor, in the permits, in the rounds and the reporting. An HVK works on the system, the risk strategy and the advisory relationship with the board, and is the level referenced when certain statutory tasks require a certified expert.

Two practical notes. First, the levels describe education and certification, not seniority: plenty of MVK profiles carry more real influence in an operation than a newly certified HVK. Second, requirements around certified involvement in testing the RI&E, and around arbodienst arrangements, change over time and depend on your company size and sector, so verify the current rules for your own situation rather than copying what a neighbouring site does. If you are weighing up the role itself, we wrote separately about how the veiligheidskundige is positioned inside Dutch organisations.

In a high-hazard environment, the role changes shape

On a Seveso or ARIE site the job stops being primarily about individual hazards and becomes about demonstrability.

You are no longer only answering "is this safe". You are answering "can we show, with documents that agree with each other, that this is controlled, that the control was tested, and that the result was acted on". That brings a vocabulary with it: the veiligheidsbeheerssysteem and its elements, the PBZO document, the veiligheidsrapport and its periodic actualisation, management of change, ageing, and the installation-transcending PDCA cycle. Supervisory attention from bodies such as the Nederlandse Arbeidsinspectie and DCMR tends to land there, on whether the management system itself functions, rather than on one valve.

The failure mode is rarely open non-compliance. It is drift. A change implemented on the installation but never carried through into the safety report. A task that formally exists but has no real follow-up. An inspection still being ticked off that says less and less about reality. An eye wash station a few weeks past date, written up as a violation.

What actually makes the work hard

Three problems show up in almost every conversation with people doing this job.

The checklist is too big for the time. A four-page round scheduled for two hours gets done the only way it can get done: on paper, from memory, with ticks. Mixing unrelated check items into one walk over a large installation is how things get missed, and everyone involved knows it.

Knowledge walks out the door. The experienced technician who explained a topic verbally three years ago has retired. Nobody wrote the reasoning down, so the question comes back as a finding instead of as a conversation.

Everything arrives in the same inbox. Risk assessments, rounds, near misses, audit findings and external observations all generate follow-up actions, and not every one of them deserves the safety team's time. Without triage, the serious-potential items compete for attention with a loose cable tie.

What to actually do in the first ninety days

If you are new in the role, or rebuilding it, this is the useful order.

1. Map where evidence lives. Not the ideal architecture, the real one: which shared drive, which spreadsheet, which inbox, which person's head.
2. Pick the top five risk scenarios and trace each one end to end, from assessment through measure, instruction, training and last test, to the action that closed the last finding. Note where the chain breaks.
3. Fix the rounds before buying anything. Split oversized checklists by installation, by discipline and by frequency, so a round fits the time it has.
4. Define triage openly. Agree with management which findings teams close themselves and which escalate, and write the threshold down.
5. Put change on the agenda. Every modification needs someone to ask whether the safety documentation follows it, before commissioning.
6. Bring one honest overview to the management review. Open actions, overdue items, recurring findings. Confronting beats reassuring.

Where software helps, and where it does not

Digital forms, inspection workflows and corrective-action tracking remove a specific kind of pain: lost actions, unreadable paper, evidence you cannot find on the day it is asked for. Tooling built for generic safety management, though, tends to stop at plan and do, which is exactly why we wrote about why generic safety management software falls short for Seveso-III companies.

What software does not do is take over the role. It does not make an organisation compliant, it does not own the risk, and it does not carry the judgement. A veiligheidskundige still decides what matters, management still decides what gets resourced, and the system only ever reflects the quality of what people put into it.

If you want to see what the check-and-act half of that looks like when it is structured instead of scattered, send us the kind of checklist you are working with today. We will walk through it with you honestly, including the parts where a tool is not the answer.