A 2025 study from APQC found that 92% of organizations do not consistently capture knowledge from soon-to-be retirees, and 58% of C-suite respondents say they are genuinely worried about it. That gap has a name: key person risk in manufacturing, and on the floor it looks like one person, one process, and one Tuesday when they don’t come in. This article looks at what that risk actually costs, why cross-training alone rarely closes it, and how plants are starting to close it without asking anyone to learn new software.
The Technician Nobody Else Can Replace
Renata Salcedo has laid up composite hulls at a boatbuilder outside Fort Lauderdale for 19 years. On a walk-through of the plant floor, she was mid-layup on a 42-foot center console, two fingers pressed into the resin along a seam, reading tack by feel before she let the crew pull the mold.
She didn’t check a chart. She told the crew to hold eleven more minutes, because the shop was running humid that morning and the batch of catalyst had come in a shade fast. Nobody else on that floor makes that call. There’s no laminated instruction for it, because the instruction would have to say “run your fingers along it and know,” and that’s not something you write down in a sentence.
That’s not a Renata problem. That’s every plant with a person like her standing at a station nobody has ever asked her to explain.
What Key Person Risk in Manufacturing Actually Costs the Shop Floor
If Renata is out for a week, the crew guesses at cure time and either scraps a hull or ships one that’s under-cured. If she retires, the plant doesn’t just lose a technician — it loses the only calibration instrument it had for that specific layup, and nobody notices until the next humid morning goes wrong.
This is what key person risk in manufacturing actually is: not a training gap, but a single point of failure with a name, a shift schedule, and an exit date nobody has put on a calendar. According to APQC’s October 2025 study, sponsored by knowledge-management vendor eGain, 92% of organizations have no consistent way to capture what people like Renata know before they leave, and 58% of the C-suite already treats this as a live problem, not a hypothetical one. That’s not a small-plant issue. That’s most plants, most of the time.
Why Cross-Training Alone Doesn’t Fix It
The standard answer is cross-training: put a second person on the layup, have them shadow Renata for a few weeks, call it done. It helps, but it doesn’t close the gap, because shadowing captures the steps, not the judgment behind them.
A shadow can learn to press two fingers into the resin. They can’t learn, in three weeks, what “a shade fast” feels like in October versus March, because that comes from years of small corrections nobody wrote down. Cross-training buys you a second pair of hands. It doesn’t buy you a second Renata.
The plants that actually reduce this risk do something more specific: they capture the reasoning behind the work, not just the sequence of steps, and they do it while the expert is still around to be interrupted with questions. We wrote about the documentation side of this gap in how experienced floor staff can hand off what they know before they leave, which covers the parts of the job that never made it into a binder in the first place.
How Operators Hand Off What They Know Without Writing a Word
The reason most capture efforts stall is that they ask the expert to become a technical writer, and Renata didn’t sign up to write documentation — she signed up to build boats.
At SenseiLab, our SOP Sprint is built around a different mechanic: operators, supervisors, and managers interact directly with our AI agent to capture how they actually do the job, in a normal conversation, with no new tool to learn. The agent asks the follow-up questions a good interviewer would ask — what changes when it’s humid, what tells you it’s ready, what would you tell the next person — and structures the answers into a Living SOP. Nobody has to sit down and write anything. A 30-day, fixed-scope engagement turns 5 to 10 of these conversations into validated procedures, trains one internal champion supervisor to keep the system running, and documents a Knowledge Capture playbook so the next Renata gets caught earlier.
Run This in Your Plant This Week
You don’t need a consultant to find out how exposed you are. This week, ask every shift supervisor two questions about each critical station:
- If the person who runs this best didn’t show up tomorrow, would the line slow down or stop? Name the person, not the role.
- Right now, today, could anyone else do that job at the same quality without help?
Tally the answers on a simple 0-1-2+ scale: how many people, besides the named expert, could actually step in. Anyone scoring a 0 is a single point of failure sitting on your floor today, whether or not it’s on anyone’s radar yet.
Frequently Asked Questions About Key Person Risk in Manufacturing
How do we capture knowledge before experienced operators retire? You capture it through conversation, not paperwork. Sit the expert down with someone, or an AI agent, whose only job is to ask how they actually do the work, and turn the answers into a structured procedure. Waiting for an exit interview is too late — the conversation has to happen while they’re still doing the job every day.
What is key person risk in a manufacturing plant? It’s any role where one specific person’s absence would stop or seriously slow production, because nobody else has the skill, judgment, or informal knowledge to cover it. It’s different from a normal staffing gap because the risk lives in one individual, not one open position.
Is key person risk the same thing as a skills gap? Not quite. A skills gap means nobody on the team has a skill. Key person risk means exactly one person does, and nothing is written down or transferred if that person leaves, gets promoted, or is simply out sick.
About the author: Diego Echenique, MBA, is CEO and co-founder of SenseiLab, with 20+ years leading Lean and Six Sigma operations across global manufacturing plants.
