If it’s faster for a person to do it, let a person do it — Approaching visual inspection through the lens of “optimizing the roles of people and machines”
ovit"We want to automate visual inspections"
This is a question we’ve been getting a lot lately.
We understand the need for automation. The president has already given the go-ahead.
But we don’t know where to start.
We’d like you to take a look at our processes and advise us on what might be feasible.
In reality, consultations almost always begin in this rather vague state.
It’s not that they want to use AI or implement the latest technology.
“First of all, I don’t even know where to start.”
I think that’s the honest truth for many people on the front lines.
Visual inspection is a process that does not add value
First, there’s something fundamental you need to keep in mind.
Visual inspection is a process that does not add value.
Just because an inspection is performed does not mean the product’s price will go up.
Nor does it mean sales will increase.
That is precisely why we cannot afford to spend a significant amount of money on visual inspections.
If we consider automation without this premise,
- Expensive equipment
- Excessive precision
- More features than necessary
We tend to end up seeking that.
As a result, we often find that even after implementing the solution, the workload on the front lines doesn’t actually decrease.
The reality is that this is a process where it’s difficult to demonstrate a return on investment
Automating visual inspections is not an investment where
"if it goes well, you'll make a profit."
What you get is mainly,
- Reduce staff
- Improve decision-making consistency
- Reduce problems
These are the defensive benefits.
If we ignore this reality
and start thinking, “If we’re going to automate, we might as well do it perfectly”
or “We might as well have machines do everything”
things quickly become precarious.
Automation can sometimes lead to a decline in productivity
When you look at the situation on the ground,
it’s not uncommon to see cases where automation has actually led to a drop in productivity.
- Every time the machine stops, someone is called
- Adjustments and recovery will take time
- In the end, someone is reviewing every single one
When it comes to this,
the number of steps increases, the effort increases, and the burden increases.
It’s putting the cart before the horse.
Much of the problem stems from the misconception that
"automation = leaving things to machines"
.
If it’s faster for a person to do it, then a person should do it
When it comes to automating visual inspections, there is a philosophy that we have consistently valued.
If it’s faster for a person to do it, then a person should do it.
Don’t force machines to take over.
Don’t aim for perfection.
Leave decisions to people when human judgment is the most rational approach.
This is not to say that I am against automation.
Quite the opposite, in fact.
The key is to "optimize the roles of people and equipment."
Automating visual inspections is not simply a matter of
"whether it can or cannot be done."
- I'll leave it to the machine from here on out
- From here on, people will see
It all comes down to how you design that division of roles.
If you try to have machines handle everything,
both costs and effort will skyrocket.
On the other hand, by dividing up the roles,
- People can focus on the decisions they need to make
- The machine silently carries out only the tasks it is good at
This results in a practical and robust process.
That’s why the first thing to look at isn’t “technology” but “processes”
When someone says, “Please take a look at our processes and advise us on what might be possible,”
we don’t immediately jump to conclusions like,
“This part can be automated,” or “This part is difficult,”
.
The first thing to look at is,
- Where is it stopped?
- Who is in trouble?
- Where am I spending my time?
Here is an overview of the entire process.
When you break it down,
"This is faster if done by a person"
"This is more reliable if left to a machine"
these points naturally become clear.
Automation is about optimization, not replacement
Automating visual inspections—
—
does not mean replacing people with machines.
It involves identifying the strengths of both people and machines,
and optimizing their respective roles.
Because visual inspection is a process that does not add value,
if approached incorrectly, it can actually reduce productivity.
That is why we believe that
"If a person can do it faster, a person should do it"
We approach automation based on this premise.
