There is a moment in nearly every Total Productive Maintenance (TPM) rollout when leadership asks: "So autonomous maintenance just means operators clean the machines?" The answer is no. The fact that this question persists explains why most autonomous maintenance (AM) programmes fail before they produce a single measurable improvement in equipment reliability.
AM is the practice of transferring specific equipment care tasks — inspection, lubrication, adjustment, cleaning, and minor repair — from maintenance technicians to the operators who run the equipment daily. The goal is not to cut maintenance labour hours. The goal is to create a condition where the person closest to the machine detects early warning signs before they become failures.
I have audited plants that rolled out AM as a compliance initiative: a checklist to complete, a form to submit, a metric to report. The difference between success and failure has nothing to do with training materials or audit sophistication. It has everything to do with whether the organisation understood that AM is a transfer of ownership, not a redistribution of chores.
What AM Actually Requires
TPM textbooks describe AM in five or seven steps depending on the framework. The specific number matters less than the progression they represent. When done honestly, that progression follows a strict sequence — one that most plants abandon by step two.
The Honest AM Progression
- 01Initial cleaning and inspectionDiagnostic, not cosmetic. Exposes hidden deterioration — oil leaks, loose fasteners, cracked hoses, contaminated sensors.
- 02Eliminate contamination sourcesEngineer the equipment and environment so maintenance requires less effort. Skip this and cleaning tasks become tedious and abandoned.
- 03Develop visual inspection standardsOperators and technicians collaborate to define what to check, how often, and what criteria indicate a problem. Built by the people who execute them.
- 04Train operators in minor maintenanceHands-on, equipment-specific training delivered by maintenance technicians who serve as coaches. The relationship shifts from 'I run it, you fix it' to partnership.
- 05Operator-led autonomous inspectionOperators inspect independently, adjust within defined limits, and feed data into the maintenance planning system.
- 06Continuous improvementStandards evolve as equipment ages and operators identify better methods. The system becomes a living process.
This progression takes twelve months to three years for a single production line, depending on equipment complexity, operator turnover, and the existing relationship between production and maintenance. Anyone promising faster results is selling something.
Where It Breaks Down
The failure modes of AM are remarkably consistent across industries. I have seen the same patterns in automotive assembly plants, food processing facilities, semiconductor fabs, and packaging operations. The methodology does not change — but the excuses do.
The Cleaning Trap
Most programmes start with initial cleaning and never progress past it. Operators are handed brushes and rags, told to restore equipment to like-new condition, and given a weekly cleaning checklist. Contamination sources are never addressed. Hard-to-reach areas are never redesigned. Inspection standards are never developed. What remains is a janitorial routine disguised as a reliability programme.
This happens because initial cleaning produces visible, photographable results that look impressive in a presentation. Eliminating contamination sources produces invisible results — the leak that stopped, the dust that no longer accumulates, the inspection that now takes three minutes instead of fifteen. Leadership rewards what they can see and neglects what they cannot.
The Competency Gap
AM asks operators to develop skills they were never hired for. Not every operator wants to learn how to read a vibration trend or adjust a chain tensioner. Plants that succeed invest heavily in the foundational knowledge that makes those skills meaningful. An operator can be taught to check a bearing temperature with an infrared gun in five minutes. Teaching that operator why bearing temperature matters, what causes it to rise, and what the consequences of ignoring it are — that takes substantially longer.
The Maintenance-Production Divide
AM requires maintenance and production to share responsibility for equipment condition. This is a profound cultural shift in organisations where these functions have spent decades in adversarial positions. Production complains that maintenance cannot keep machines running; maintenance complains that production abuses the equipment and ignores warnings.
Handing inspection tasks to operators is perceived by maintenance as a threat to their expertise and by production as an intrusion into their time. Both sides are partially right. Without active management of this transition — role clarification, cross-functional training, a shared scorecard — the two departments will quietly undermine the programme until it collapses.
The Audit Illusion
Many organisations attempt to enforce AM through layered audits. Supervisors audit operators. Managers audit supervisors. Corporate auditors audit the plant. The findings generate corrective actions, which generate more audits. The system becomes self-sustaining but produces no actual improvement in equipment reliability.
Audits measure compliance, not competence. An operator who completes a checklist perfectly without understanding what they are checking for has complied with the standard without achieving its intent. The inspection data shows a 99% completion rate while equipment failures remain unchanged.
What Success Looks Like
I worked with a tier-one automotive supplier that implemented AM on their stamping presses over eighteen months. The results were not instantaneous, but they were measurable and sustained.
In the first three months, initial cleaning revealed over forty tagged issues — oil leaks from hydraulic fittings, cracked guards on flywheel enclosures, worn vibration dampers, contaminated lubrication lines. These were not new problems. They had been visible for months, but no operator had been asked to look and no maintenance technician had been assigned to inspect behind the guards.
By month six, the team had eliminated twelve contamination sources, redesigned access panels on three press stations, and developed one-page visual inspection standards — photographs of the component, the inspection method, the acceptance criteria, and the action to take if out of range. They were laminated and attached to the presses at the point of use.
| Metric | Baseline | Month 18 |
|---|---|---|
| Unplanned downtime | Baseline year | -34% |
| Mean time between failures (critical presses) | 11 days | 19 days |
| Operator turnover | Historical average | Significantly reduced |
By month twelve, operators were conducting daily five-minute inspections and weekly fifteen-minute deep checks without supervision. Maintenance technicians had shifted to predictive work — oil analysis, vibration monitoring, thermography — instead of reacting to breakdowns. By month eighteen, the programme had expanded to welding and assembly. The maintenance department had reorganised into a planning-focused structure with two technicians dedicated to AM coaching.
None of this happened because someone bought a TPM software package or hired a consultant for a two-day workshop. It happened because the plant manager committed to the progression, the maintenance manager accepted a changed role, and the production supervisor allowed operators the time to do the work properly.
Common Objections
"We don't have time for operators to do maintenance tasks." This is the most frequent and the most legitimate objection. AM requires fifteen to thirty minutes per shift. The question is whether that time is recovered through reduced downtime. In the stamping plant example, thirty minutes of inspection time was recovered within the first month through faster changeovers and fewer mid-shift stoppages. If your equipment is so unstable that thirty minutes of inspection cannot be recovered through improved reliability, you have a far bigger problem than AM.
"Our operators don't have the skills." This is often true and entirely addressable through structured training. But it raises a deeper question: do you want operators who understand the equipment they operate, or do you want operators who are interchangeable with any newly hired worker? If the latter, AM is not for you. If the former, the training investment is justified.
"Our maintenance team won't support it." This objection usually masks a power dynamic that needs to be addressed openly. Technicians who view their expertise as job security will naturally resist sharing it. This resistance dissolves when maintenance leadership redefines the technician role from "the person who fixes things when they break" to "the person who ensures things don't break, with operators as partners." That redefinition requires training, coaching skills, and a compensation structure that rewards prevention over reaction.
The Real Transfer
Equipment reliability is not a maintenance function. It is a shared responsibility that depends on the person running the machine understanding it well enough to feel when something is wrong, look for evidence of developing problems, and take action before failure occurs.
AM cannot be mandated. It can only be developed through training, trust, and time.
The plants that succeed build capability before they assign tasks. The plants that fail skip the capability-building, go straight to task assignment, and wonder why their operators fill out inspection forms without actually inspecting anything.
Start with one line, one team, and one equipment type. Invest in training before you invest in checklists. Address contamination sources before you schedule cleaning rounds. Let maintenance technicians coach operators before you hand over responsibility. Measure results in downtime reduction, not in audit scores.
The goal is not cleaner machines. The goal is operators who know their equipment well enough to prevent failures before they happen. Everything else is implementation detail.