Who Owns Safety When Nobody Has It in Their Title?
Who is responsible for machine safety at an integrator with no safety engineer? A practical guide for the controls engineer or PM who just inherited it.
)
Who is responsible for machine safety in an integration project? Legally and contractually, the company that builds and delivers the system. Practically, inside most integrators, it's whoever happens to be standing closest when the customer asks for the risk assessment.
Sometimes that's the controls engineer, because safety functions live in the PLC and the PLC is theirs. Sometimes it's the project manager, because the safety case is a deliverable and deliverables are theirs. Sometimes it's the owner of the company, three weeks before commissioning, reading a procurement checklist that mentions documents nobody scoped.
Most integrators have no one with "safety" in their title, and hiring someone isn't a quick fix even when the budget exists. Experienced safety engineers command $180K to $250K and more, they're rare, and the market is producing them slower than the industry needs them. So the role gets absorbed instead of assigned, and the person absorbing it is expected to produce work in a discipline they were never trained in, on a deadline set by someone else's contract.
If that person is you, this post is your starting map. If you run the company, it's an argument for making the assignment deliberate instead of accidental.
What does the safety owner actually have to produce?
Strip away the intimidating vocabulary and the core deliverables for a typical integration project are countable:
A risk assessment of the assembled system, following ISO 12100 methodology: identify hazards, evaluate risks, document how each one gets reduced to an acceptable level.
Safety requirements derived from that assessment, including the required performance levels for safety functions under ISO 13849.
Verification that the safety functions meet those levels, which is the reliability math connecting components to performance claims.
Validation records proving the system as built behaves as the safety design says.
The assembled safety case, with the traceability connecting all of the above, handed to the customer at commissioning.
The list is finite. What makes it feel infinite is the blank page. The most significant barrier for a first-time safety owner is starting from nothing: a new standard, an unfamiliar method, and no institutional template, on top of a day job. Meanwhile the industry default toolchain, spreadsheets for the hazard log, a desktop calculator for reliability, documents for everything else, none of which talk to each other, means even experienced practitioners lose 60% of their safety time to documentation mechanics rather than judgment.
How do you assign safety ownership properly?
Name the owner at project kickoff. The same meeting that assigns the electrical design and the PLC work assigns the safety case. Safety engineering is scoped work like anything else. Unscoped, it becomes a crisis with a commissioning date.
Scope it into the contract. The risk assessment, the safeguarding design, and the validation records are line items with hours attached. Integrators who price this work win twice: the project carries its real cost, and the customer sees a partner who knows what the deliverable is before being asked.
Give the owner authority to match the accountability. A safety owner who can't influence cell layout or component selection is a documentation clerk with liability. The assignment has to come with a real voice in design reviews while changes are still cheap.
Accept that the owner will not be an expert, and structure around it. This is the decision that separates functional programs from fictional ones. You're not going to turn a controls engineer into a 15-year safety veteran by Q4. You can give them a process that carries the expertise for them.
Can a non-specialist really carry the safety case?
Yes, with structure doing the heavy lifting. This is precisely what safety democratization means: the methodology encoded into guided workflows, so the process itself knows what comes next, what each step requires, and what the output has to look like.
That's the design premise of ASAP. Guided workflows replace blank-page safety documentation with structured, traceable outputs. Pre-configured templates for autonomous mobile robots, collaborative robots, and industrial machinery mean the hazard patterns of common applications arrive already mapped, and your team's job becomes evaluating and adapting them for the specific cell, which is engineering judgment they already have. The platform brings proven methodology from Amazon to teams without resident specialists, which matters because it was built over 5 years inside Amazon Robotics and now protects more than 1 million robots in 24/7 operation.
The division of labor that emerges is the sustainable one: the accidental safety owner runs the process day to day, documents decisions as they're made, and keeps traceability current, while scarce senior expertise, whether an internal veteran or an external reviewer, reviews and decides at the checkpoints instead of doing everything. Guided workflows turn months of coordination into weeks of focused engineering, and the same structure means the second project inherits the first one's framework instead of starting over.
One more thing the structure buys: an exit from the single-expert bottleneck before it forms. Companies that concentrate safety knowledge in one head rediscover the problem every time that head is on vacation, on another project, or on the way out. When the process carries the expertise, it survives personnel changes. When one person is the process, it doesn't.
If the safety case for your next cell currently belongs to "whoever gets asked," it's worth an hour to fix that before the project starts. Talk to Fennec about giving your team the process.
Frequently asked questions
Who is responsible for machine safety in a robot integration project?
The integrator is responsible for the safety of the assembled system, including the risk assessment, safeguarding, and validation. Robot and component manufacturers answer for their products as supplied. Inside the integrator, that corporate responsibility should be assigned to a named individual at project kickoff, with scoped hours and design authority.
Do we need to hire a safety engineer?
Not necessarily, and for many integrators it isn't realistic: experienced safety engineers cost $180K to $250K or more and are hard to find. A structured, guided safety process lets existing engineers carry the day-to-day work credibly, with senior or external expertise concentrated at review points.
What should the safety owner do first on a new project?
Start the risk assessment at kickoff, before the cell layout freezes. Hazards identified early shape design decisions cheaply. Hazards identified at commissioning become rework, and rework from late findings is where integration margins go to die.
Can a project manager own the safety case?
Ownership of coordination, yes: schedule, deliverables, and traceability discipline suit a PM well. The engineering judgments inside the case, hazard evaluation, performance level decisions, and validation signoff, need engineering competence, whether from the team, a senior reviewer, or both.
What happens if nobody owns safety on a project?
The deliverables still come due; they just arrive as a surprise. The typical sequence is a customer request for the risk assessment near commissioning, an unbudgeted scramble to reconstruct decisions after the fact, and a safety case that documents what the team hopes they built. Naming an owner at kickoff costs a meeting. Not naming one costs the end of the project.
)