Part four of Monday Morning Plant Manager, a series about the problems that meet you at the door of the plant, and what they mean for the rest of the business.
Monday morning, and the extruder on line 4 is running out of spec again. Two technicians have been on it since six. At 8:40 somebody calls Ray, who retired in March, and asks what he used to do when it behaved like this. He talks them through it in about four minutes, from his kitchen. Nobody writes it down, because Ray answered the phone. Manufacturing knowledge transfer can wait, until the morning he doesn’t.
Every plant has a Ray, and usually several. What they hold is a real operating asset: the failure modes that don’t appear in any manual, the setting that works on the second press but not the first, or the sound a bearing makes three weeks before it goes. That asset has never appeared on an asset register, carries no maintenance schedule, and depreciates on a calendar nobody in operations controls — the relentless march of time.
The retirement itself is usually visible years ahead. The harder problem is that the plant has organized itself around knowledge it has never written down, and the cost of that arrangement only becomes legible once the person holding it stops answering the phone.
The signals knowledge is concentrated in too few heads
- The experienced bench is thinning. Operators, technicians, and maintenance specialists with decades on the floor are retiring in a cluster, and the pipeline behind them is thinner than the one that produced them.
- Troubleshooting lives in people rather than systems. The diagnostic path for a recurring fault exists in somebody’s memory, and the written procedure everybody knows to ignore covers the version of the machine that was installed in 2009.
- Onboarding keeps getting longer. New hires take longer to reach competence than they used to, because the equipment is more complex and the person who would have trained them is doing three jobs.
- Production depends on a handful of people. There are names that, when they appear on the vacation schedule, change what the plant is willing to run that week.
- The same problems need the same experts. A fault that recurs every few months and is solved by the same person every time is a fault the organization has never actually learned from.
- Consistency varies by shift, line, or site. The same product, made to the same spec, comes out slightly different depending on who’s running it, which means the real process is the one in people’s habits.
Read down that list and a pattern emerges. The plant has an operating method, and most of it is in the heads of the people who do the operations.
Why undocumented expertise is a business risk
Operational knowledge behaves like any other critical asset: when it becomes scarce, everything downstream slows. Learning curves lengthen, productivity dips during the handover that never formally happened, processes drift apart across shifts, and execution risk rises on exactly the work that needs consistency most.
The skilled labor shortage in manufacturing only exacerbates this problem. Capgemini’s 2026 research on reindustrialization in Europe and the US finds that shortages of skilled industrial and digital talent are actively limiting the scale at which manufacturers can deploy automation, AI, and digital twins, even in tech-intensive new facilities. A plant can buy capability faster than it can staff it.
Factory leadership has noticed. In Accenture’s global survey of 552 factory managers, knowledge management, embedding data analytics into daily workflows, and enabling data-driven decision-making rank as their top focus areas, with the workforce shifting over time from working in production toward working for production: oversight, decision-making, and optimization instead of manual execution. Judgment becomes the job, and judgment is the part currently sitting in Ray’s head.
Knowledge transfer before the expertise leaves
Start with the failures. Full process documentation is a project that never finishes. The recurring faults are a finite list, and they’re where the undocumented knowledge actually sits. Ask the experienced people which problems they get called about, and capture those first.
Capture in the format the work happens in. Expecting a technician to write a manual produces nothing. A short video taken at the machine, an annotated photo, or a recorded walkthrough during an actual repair takes minutes and preserves the detail a written procedure loses. The test is whether the next person can find it and use it at the machine, in the moment they need it.
Put the knowledge where the work is. Information filed somewhere a technician can’t reach at 2 a.m. with gloves on has been archived rather than shared. IDC projects that by 2027, more than half of manufacturers will use AI-enabled knowledge management tools to reskill and upskill their workforce and to support collaboration across their operations. Those tools retrieve what has been captured, so the capture habit comes first.
Standardize the process before automating it. Where the same job is done three ways on three shifts, the first gain comes from agreeing on one way, not from buying a system. Standard work also makes the eventual automation far easier to specify, because there’s a defined process to encode.
Aim automation at the dependence. Automating the repeatable, physically demanding, and error-prone portions of a job frees experienced people for the diagnostic work only they can do. Automation applied to the judgment itself tends to disappoint, because the judgment is the part that was never written down.
Train for the tools you’ve installed. BCG’s global survey of more than 10,600 employees found that regular use of AI tools is sharply higher among people who receive at least five hours of training with in-person coaching, and that positive sentiment rises from 15 percent to 55 percent where leadership support is strong, though only about a quarter of frontline employees report getting that support. A system nobody was trained to trust gets bypassed.
Where to start when the bench is already thinning
Begin where the dependency is sharpest. List the people whose absence changes what the plant will run, and work outward from there: what they know, what breaks when they’re away, and what a competent replacement would need in order to handle it. The list is short, specific, and usually uncomfortable reading.
Sequence by exposure. The person retiring in eight months matters more than the one retiring in five years, and the fault that recurs monthly matters more than the one that struck once in 2021.
Then keep it running. Manufacturing knowledge transfer can’t be a one-time exercise. Building it into the shift, the maintenance routine, and the handover is what keeps it current, and what turns one person’s expertise into something the whole operation can draw on.
Turning expertise into an asset the plant owns
This is the work Eclipse does with manufacturers every day: standardizing processes so they can be taught and repeated, designing automation that removes dependence on individual availability while keeping skilled people on the problems that need them, and building systems that put operating knowledge at the machine where the work happens. Our Advanced Engineering Services help identify where that investment returns most, and the automation lifecycle that follows includes the training and support that turn a new system into one your team trusts.
Ray will keep answering the phone for a while. The question is what the plant knows by the time he stops.
More from Monday Morning Plant Manager
Monday Morning Plant Manager takes the problems waiting on the schedule board and breaks down what they really mean for your business right now, not in a hazy half-promised future. Also in the series: whether the plant can carry the company’s growth plans, what delaying an automation decision is already costing you, and why automation plans stall between pilot and production.
FAQs
Frequently asked questions
How do we capture troubleshooting knowledge without asking experts to write documentation?
Record it instead of writing it. A phone video taken during an actual repair, with the technician narrating what they’re checking and why, captures more usable detail in five minutes than a written procedure captures in an afternoon, and it doesn’t ask a hands-on person to become a technical writer. Index by symptom, because symptoms are how the next person will search.
What should we document first if we can only do a little?
The faults that recur and the people who get called about them. Ask three or four experienced staff which problems they’re phoned about most, and you’ll have a prioritized list within a day. That list covers the majority of unplanned interruptions in most plants, and it’s finite in a way that full process documentation never is.
How do we shorten onboarding when the manufacturing skills gap makes experienced hires scarce?
Separate what a new hire needs to know from what they need to be able to find. Competence on the routine work, plus reliable access to guidance on the exceptions, gets someone productive far sooner than trying to transfer everything up front. For the diagnostic portion, structured shadowing on real faults beats classroom time.
Can automation replace the expertise we’re losing?
Not the judgment. Automation is very good at removing the repetitive and physically demanding work that consumes experienced people’s time, and at enforcing consistency that currently depends on who’s on shift. The diagnostic intuition built over decades has to be captured and taught. Treating automation as a substitute for that tends to produce a plant that runs well until something unusual happens. Which it always does.
How do we get consistency across shifts and sites?
Define the standard, make it visible at the point of work, and measure adherence instead of assuming it. Where shifts differ, the gap is usually a legitimate workaround somebody developed for a real problem, so the first step is finding out why the variation exists. Learn more about Eclipse’s smart manufacturing solutions.
Explore the possibilities
Ready to make your plant’s expertise easier to keep and easier to scale? Book a discovery call to learn how Eclipse Automation supports process standardization, workforce-enabling automation, and knowledge transfer in manufacturing.
Related articles and insights
-
Monday Morning Plant Manager: What happens when decades of factory knowledge walk out the door?
Every plant has experienced people whose know-how has never been written down, and retirements are thinning… Read More
-
Aerospace precision tooling for critical machining
Aerospace precision machining depends on tooling engineered for accuracy and repeatability. Learn how Eclipse Automation designs… Read More
