It is a Tuesday morning. I am at my desk with coffee, looking at a root cause analysis the dashboard pulled together overnight from an inspection finding the day before. I am not the EHS director at the site in question — but I built the tool, and I run a handful of synthetic inspections through it every week to see what comes out the other side. The RCA on screen is for a pinch-point near-miss at a screen plant. Five whys. Three corrective actions. Two citations to the relevant standard. Audit-grade.
It is also, I realize, most of a toolbox talk I had not been planning to write.
The RCA's second job
For most of the time I have been around EHS programs, the root cause analysis has been treated as a back-office artifact. You produce one because the regulator expects to see one. It lives in a binder, or in a folder on a network drive, or — at the better-run sites — in a slot of the inspection record that the next auditor will flip to. Its primary consumer is the inspector who shows up three months later. Its job is to demonstrate that the site understood what went wrong and what changed.
That framing is correct. It is also incomplete.
The RCA's first consumer is the auditor. Its second consumer — the one most platforms in this space have never thought to optimize for — is the crew that is going to clock in tomorrow morning.
What I noticed
The structure of a good RCA looks remarkably like the structure of a good toolbox talk. There is a setup: this is what happened. There is a causal chain that goes one layer deeper at each step until it lands on something a person can actually change. There are corrective actions: this is what we are doing about it. And there are citations: this is why those actions are the right ones, anchored to a specific section of the standard that the crew is already trained against.
Reframe that paragraph for a five-minute pre-shift huddle and you have your morning's safety talk. Most of the work from here is translation — into the voice the crew actually talks in, with the names of the people and the equipment involved on this site — rather than authoring. The thinking is already done.
I noticed this somewhere around the third or fourth time I caught myself opening an RCA, then opening a fresh document, then typing some version of the same content into the fresh document in a friendlier register. The loop was sitting right there and I had not seen it.
What I had been getting wrong
For longer than I would like to admit, I had been treating the RCA artifact and the toolbox talk artifact as two separate workflows. Two different surfaces. Two different efforts. The toolbox talk was a meeting agenda item owned by the shift supervisor, lived in its own form, its own folder. The RCA was an inspection deliverable owned by the EHS lead, lived next to the finding.
The work of writing the second one already existed inside the first. I had built two doors into the same room without realizing it. The actual unlock — the thing that changed how I think about both — was just stopping and noticing that one of the documents was already inside the other.
What the dashboard deliberately does not do
A note on what it does not do here, because this is the part where it would be easy to get glib.
It does not write the toolbox talk for you. The translation from the audit-grade RCA into the pre-shift huddle script is yours. The crew's voice is yours. The site context — the names, the recent near-misses, the rapport — is yours. The dashboard produces the source material; you produce the talk. That distinction matters, because the moment a tool starts authoring the safety conversation, the safety conversation stops being the supervisor's.
It also does not pretend that the regulator-grade output is the right tone for the pre-shift huddle. Citations are useful for the auditor; they are noise in front of a crew who already knows the standard. The RCA artifact and the toolbox talk are two different documents with two different audiences, and we kept them that way on purpose. What changed is that one is now an obvious input to the other.
The loops nobody pitched
The lesson I am taking away from this is that the most useful thing a platform can do is sometimes show you a loop you had been walking around. The RCA-to-toolbox-talk loop was not a feature anyone asked for. It is a side effect of having the structured output sitting in a place where I would see it the next morning and notice what it already contained.
I would be interested to hear from other EHS folks about loops like this in your own workflow. Inspection findings into training records. Permit data into route planning. Hours-worked patterns into staffing decisions. The connections nobody pitched you — the ones you spotted yourself because the artifacts happened to be sitting next to each other. The honest answer to "what is the platform good for" is sometimes "it lets the operator see the next loop on their own."
Get in touch
Glad to be writing this one. Next post when something is worth saying.
— Jeffrey