Redesigning a hot work permit system from scratch — and what happened when I treated a compliance problem like a product problem.
Subcontractors on our sites were skipping hot work permits at an unacceptable rate. More enforcement wasn't working. I approached it as a design problem instead: user research, a stripped-down MVP, a two-site pilot, iteration. The result was 85% voluntary adoption with zero enforcement, a 66% reduction in process time, and $320K in Year 1 ROI. More importantly, it changed how I think about why safety programs fail.
Hot work — welding, cutting, grinding, anything that generates heat or sparks — is one of the leading causes of construction fires. A hot work permit is the industry's answer: before you start, document what you're doing, where you're doing it, what precautions you're taking. If something goes wrong, there's a record. More importantly, the act of filling it out is supposed to slow you down enough to actually think about the hazard.
The problem was that subcontractors were skipping them. Not occasionally. Routinely.
The standard interpretation is negligence or willful non-compliance. The standard response is more enforcement: spot checks, written warnings, suspensions for repeat offenders. We had tried those things. They produced temporary compliance on the sites where enforcement was visible, and no change anywhere else.
I spent a few weeks actually talking to the subcontractors who were skipping permits — not to discipline them, but to understand what was happening. Nobody thought the permits were a bad idea. They understood why they existed. They weren't skipping them out of laziness or disregard for safety.
They were skipping them because the process took 45 minutes, and most of the hot work they were doing took 20 minutes. The math didn't work. From their perspective, a process that took more than twice as long as the task itself wasn't a safety measure — it was a broken process that someone else was enforcing on them.
They were right.
Once I stopped looking at this as a compliance problem and started looking at it as a usability problem, the path forward became clearer. The permit process was slow for three reasons — none of which were actual safety requirements:
None of those were safety requirements. They were accumulated process friction. The core safety requirement — document the hazard, confirm the precautions, get it on record — could theoretically happen in under ten minutes. We were carrying forty minutes of overhead on top of it.
I redesigned the permit from scratch with a specific constraint: a field worker doing routine hot work should be able to complete the permit in ten minutes or less, without having to find a supervisor until the final sign-off.
I piloted on two sites for six weeks before rolling out broadly. Those sites were chosen specifically because they had the worst permit compliance records — if it worked there, it would work anywhere.
The pilot sites went from roughly 40% permit compliance to over 80% within the first three weeks — and compliance kept climbing as the process became familiar.
When people aren't following a safety process, the instinct is to enforce harder. Sometimes that's right. But it's worth asking, before you escalate enforcement, whether the process itself is the problem. Subcontractors on our sites were rational people making a rational calculation: this process takes more time than it's worth. They weren't wrong. Enforcing it harder would have produced resentment and more sophisticated workarounds, not safety.
I also learned something about how safety programs accumulate dysfunction over time. Every field that got added to the permit form was added for a reason — usually after an incident, an audit finding, or a regulatory change. Nobody was there to remove things when the reason went away. The permit had become a monument to past problems rather than a tool for current ones. That's a maintenance problem, and it affects almost every long-running compliance program I've seen.
The pattern I keep running into — across orientation systems, asset tracking, permits, and everything else I've built — is the same: safety and compliance programs get designed for the compliance officer, not the person in the field. The documentation is the product. The safety outcome is secondary. That's not a cynical observation. It's a design problem with a design solution.
If you're working on EHS technology or evaluating tools for field operations and seeing adoption problems from people who genuinely want to do the right thing — that's usually diagnosable and fixable. I do consulting work in this space.