Autonomous
Maintenance — Jishu Hozen: When the Operator Becomes the Doctor of Their
Machine and Your Maintenance Gains Thousands of Eyes
Imagine a factory where every operator knows their machine so well
that they catch an anomaly before it becomes a breakdown. Not because they
have superpowers — but because they were taught how. This isn’t utopia.
This is Autonomous Maintenance.
The First
Encounter with the Machine That Doesn’t Speak
It was in an automotive plant in central Slovakia where I first saw
Autonomous Maintenance in action. A young operator — Martin, twenty-three
years old, only six months on the job — stopped by a CNC milling machine,
placed his hand on the cover, closed his eyes, and said: “The Y-axis
bearing needs to be replaced within two weeks. It started squeaking.”
The maintenance technician standing next to me smiled. “You’re right,”
he said. “The ultrasonic inspection confirmed it yesterday. But he heard
it before we did.”
I asked Martin how he knew. “When you sit here eight hours a day, five
days a week, that machine becomes your partner. I know what it sounds like
when it’s happy. And I know what it sounds like when something’s
bothering it.”
That’s Autonomous Maintenance. Not magic. Not talent.
A systematic process that transforms the operator from a
passive observer into an active guardian of their equipment.
What Is Autonomous Maintenance?
Autonomous Maintenance (AM), or in Japanese Jishu Hozen
(自主保全), is one of the eight pillars of Total Productive Maintenance
(TPM). Its essence is simple yet revolutionary:
Transfer basic maintenance from the hands of technicians into the
hands of operators — and create a partnership where both do what they’re
best at.
This doesn’t mean the operator replaces the technician. It means the
operator performs partial maintenance tasks — cleaning,
lubricating, inspecting, tightening fasteners, detecting anomalies —
while maintenance focuses on strategic, complex, and preventive
activities.
Be honest: how many times do you call maintenance for something the
operator could resolve in three minutes? Lubricating guide rails. Cleaning
a filter. Replacing an extraction hose. Checking air pressure. These
“small” tasks consume tens of thousands of hours per year — and
simultaneously create a dependency that teaches the operator one thing:
“When something’s not right, it’s not my problem.”
AM destroys this pattern. And replaces it with a new one:
“This is my machine. I know it. I take care of it. And when
something’s not right, I’m the first to know.”
Seven Steps — The
Path from Cleaning to Ownership
Autonomous Maintenance isn’t a single workshop or a banner on the wall.
It’s a seven-step journey that takes months — sometimes
years. And every step is irreplaceable.
Step 1: Initial Cleaning
The beginning is provocative. Take a team of operators, give them rags,
brushes, cleaning agents, and say: Clean your machine like never
before.
But this isn’t ordinary cleaning. Every speck of dust, every drop of
oil, every loose screw becomes information. Operators
document everything they find:
- Oil leaking from seals
- Dust accumulated on sensors
- Damaged cables
- Missing guards
- Loose connections
This step accomplishes two things at once: it cleans the machine and
opens the operator’s eyes. For the first time, they see
what’s hiding under the covers. For the first time, they realize that the
“minor oil leak” they’ve been ignoring for months is actually a
risk.
The rule: Cleaning is inspection. Inspection is
diagnosis. Diagnosis is prevention.
Step
2: Eliminating Sources of Contamination and Hard-to-Reach Areas
Initial cleaning reveals problems. Step 2 asks: Why will the
machine get dirty again?
The team identifies sources of contamination — places where oil leaks,
dust enters, chips fall — and proposes countermeasures.
At the same time, they map areas that are difficult to clean or inspect
and make them more accessible.
Example: A CNC machine had a problem with chips getting into the guide
rails. The solution? A simple guard made from durable plastic that the
operator designed and maintenance installed. Payback? Three days.
Step 3:
Standards for Cleaning, Lubrication, and Inspection
Now that the machine is clean and contamination sources are addressed,
it needs to be maintained. Standards are created:
- What gets cleaned, how it’s cleaned, how often
- What gets lubricated, with what lubricant, in what quantity
- What gets inspected, what the criteria are, what to do when there’s a
deviation
These standards aren’t lengthy documents. They’re visual,
single-page, right at the machine. The operator has them within
arm’s reach — and an inspection takes five minutes, not fifty.
Step 4: General Equipment
Inspection
This is where the transformation from “cleaner” to “diagnostician”
begins. Operators learn basic technical skills:
- How to check pressure in a pneumatic system
- How to recognize abnormal vibration
- How to read a machine’s thermal profile
- How to identify tool wear
- How to check voltage and current
This step requires training. And this is where many
companies fail — they send operators to training but don’t give them time
to learn. AM requires an investment in people. Not a financial one — but a
time investment. And it comes back a hundredfold.
Step 5: Autonomous Inspection
Step 5 is the moment when the operator takes ownership
of routine inspections. Not because a manager ordered it — but because
they understand why they’re doing it.
Autonomous inspection means:
- The operator performs daily inspections according to the standard
independently - The operator records results independently
- The operator identifies anomalies and reports them independently
- The operator performs basic maintenance independently (lubrication,
cleaning, minor adjustments)
And — this is crucial — the operator collaborates with
maintenance on solving problems that exceed their competence.
Step 6: Standardization
All habits and procedures from previous steps are
standardized. This creates:
- Single-page standards at the machine
- Visual markings (where to lubricate, where to inspect)
- Checklists integrated into the daily routine
- An anomaly reporting system
Standardization isn’t bureaucracy. It’s a guarantee
that what one operator does well, their colleague on the night shift will
do well too.
Step 7: Fully Autonomous
Maintenance
The final step is the goal: the operator owns their
machine. Not just operates it — but understands it, cares for it,
communicates with it. This isn’t anthropomorphism. It’s
professional intimacy that every experienced craftsman
knows.
At this stage:
- The operator can anticipate problems
- The operator proposes improvements
- The operator trains new colleagues
- Maintenance focuses on prevention and optimization, not firefighting
Why Most
Companies Fail at Autonomous Maintenance
I’ll be honest with you. I’ve implemented AM in dozens of companies and
I’ve seen three failure patterns that repeat like a scratched vinyl
record:
1. “We Don’t Have Time for This”
The most common excuse. And the most nonsensical one. The operator
spends eight hours a day at the machine. Adding a five-minute inspection
reduces machine availability by 1%. But a caught anomaly that prevents a
two-hour breakdown? That’s a 2,400% return on investment.
Reality: Companies that implement AM increase OEE by
10–25% in the first year. Not because machines run faster — but because
they break down less.
2. “That’s Not the Operator’s Job”
A cultural barrier. The operator “just operates.” Maintenance “just
repairs.” Management “just manages.” These silos are killers of
excellence.
AM doesn’t just solve a technical problem. It solves a
cultural problem. And culture changes slowly, deliberately,
and with leadership that leads by example.
3. “We Tried It, but It Never
Took Off”
Step 1 alone — initial cleaning — requires discipline. Most companies
clean once, take photos, put them in a presentation, and forget about it.
But AM isn’t an event. It’s a process. And a process
requires systematic support:
- Regular audits of AM activities
- Visual boards at the machines
- Team meetings (short, frequent)
- Recognition and rewarding of progress
- Patience — results come in 3–6 months
How to Start — A Practical Guide
If you’re reading this article and feel like this could work for you,
here’s a simple plan to begin:
Weeks 1–2: Selecting a Pilot
Machine
Don’t pick the worst machine. Pick an average one —
with a team that’s open and willing. The pilot must succeed so it can
serve as a reference.
Weeks 3–4: Initial Cleaning
Two days of intensive cleaning. Before and after photos. Documentation
of all findings. And — most importantly — a shared
experience that unites the team.
Weeks 5–8: Standards and Training
Create simple standards. Train operators. Not just “how” — but also
“why.” When an operator understands the reason, they follow the standard
even when no one is watching.
Weeks 9–12: First Results
Measure. Compare. Celebrate. Even small progress is progress. Reducing
minor breakdowns by 30% in one quarter is tangible proof that AM
works.
Week 13+: Expansion
Transfer the success from the pilot to other machines. Carry the
experience from one line to the entire factory. AM ambassadors
— operators from the pilot team — become trainers for the rest.
Numbers That Don’t Speak —
They Scream
Tell me one number that will convince you. Here it is:
Average share of breakdowns caused by inadequate basic
maintenance: 40–60%.
This means that nearly half of all unplanned downtime could be prevented
through basic care — cleaning, lubricating, inspecting, tightening.
Things that cost little and take little time.
And another number: companies with implemented AM achieve a
15–30% reduction in unplanned downtime in the first year. That’s
not a hypothesis. That’s data from thousands of implementations
worldwide.
Autonomous
Maintenance and Quality — The Invisible Connection
You might ask: “This is about maintenance. What does it have to do with
quality?”
The answer is simple: everything.
Damaged guide rails = inaccurate machining = scrap. Contaminated sensor
= incorrect feedback = defect. Loose screw = vibration = poor surface
quality. Leaking oil = contamination = customer complaint.
Every machine anomaly is a potential cause of poor quality.
And the operator who catches it in time isn’t just a maintenance tech —
they’re the first line of quality defense.
The Japanese Philosophy Behind
AM
The Japanese word Jishu (自主) means “autonomous,” “one’s
own,” “independent.” Hozen (保全) means “preservation,”
“conservation,” “maintaining.” Together: self-preservation.
This isn’t a random name. It’s a philosophy: equipment is a living
organism that requires daily care. And this care must come from
the person who is in daily contact with it — not from a specialist who
visits once a month.
It’s like a garden. You can have the best gardener in the world — but
if they only spend an hour a week with the plants, they won’t notice when
they need water. The person who sees them every day will know.
Maintenance
4.0 — Autonomous Maintenance in the Digital Age
How does AM change in the era of Industry 4.0? Fundamentally — and at
the same time, not at all.
Fundamentally in that the operator now has access to new
tools:
- IoT sensors — automatic measurement of vibrations,
temperature, pressure - Mobile apps — checklists on a tablet, not on paper
- AR glasses — visual instructions directly in the field
of view - Dashboards — real-time data on equipment condition
But not at all in that the human factor remains crucial.
A sensor will tell you that vibrations exceeded the limit. But only the
operator knows that “today the machine sounds a bit different, even though
the numbers are still within range.” Technology amplifies human senses.
It doesn’t replace them.
Conclusion:
The Machine That Speaks — and the Human Who Listens
Autonomous Maintenance isn’t about maintenance. It’s about the
relationship between human and machine. About the fact
that the operator isn’t just “the one who pushes buttons” — but a
guardian, diagnostician, and partner of their equipment.
I know what you’re thinking: “That’s a nice theory, but it won’t work
here.” Maybe. But until you try — really, systematically, with patience
— you won’t know.
And remember: that operator who heard the squeaking bearing? Three months
later he proposed an improvement that reduced cycle time by 12%. Because
when a person starts listening to their machine, the machine
starts talking to them.
And what it tells you could save your line — and your budget.
Peter Staško is a Quality Architect with 25+ years of experience in
automotive, manufacturing, and continuous improvement. He has implemented
Autonomous Maintenance across dozens of production lines and helps
companies transform operators into partners of their machines. He believes
the best maintenance system is one where every machine has its own
doctor.
