How New Technologies Evolve

Standardized Work in Industry 4.0

A guide to how computer vision, AI and video analytics
can help drive continuous improvement
Introduction

The manufacturing world is moving toward Regardless of the pros and cons of robots
the Industry 4.0 vision: automating versus humans, the world is still far from full
manufacturing processes to realize automation. Today 72% of the world's
productivity levels and quality control assembly tasks are still done manually. These
unachievable by human hands. This vision is tasks represent 345 million workers
based on real challenges that human workers worldwide. While there may be a distant
present: variability in assembly processes that future where automation has a stronger
can lead to decreased throughput and presence on the line, our greatest asset today
increased quality issues; labor shortages; continues to be people, and standardized
supply chain disruptions; the list goes on. On work is our most powerful tactic to minimize
its surface, automation provides an attractive variability.
alternative to people.
All manufacturing companies have standard
But the vision of total automation neglects the operating procedures (SOP) of various kinds.
very real drawbacks of machines on the The goal is to put guardrails around people so
factory floor. For starters, automation is they follow some type of process that has
expensive and inflexible. First, the capital been determined to be efficient and effective.
costs of new equipment are depreciated over For repetitive manual work, the SOP is
many years, and if demand goes down for standardized work.
products from that automated line, the
company still pays for the machine, whereas Standardized work, which predates lean by
people can be redeployed. more than 50 years, defines the steps, order
and timing of a given task in an attempt to
Second, even the best robots do not have the control variability. As you’ll see in the next
dexterity and sensory capabilities of humans. section, in the early days going back to
Frederick Taylor, the father of industrial
Third, when there are even small changes in engineering, standardized work was seen as
the work, reprogramming robots can be an the purview of engineers who defined the
onerous task. work, while managers commanded and
controlled the workers who were expected to
Finally, and perhaps most important, people follow it without deviation.
have the capability for continuous
improvement, sensing and creatively solving Toyota redefined the way standardized work
problems, that robots do not. was used with the goal of empowering those
closest to the work, line associates and
supervisors, to use standardized work as the
basis for improvement. In fact, many will tell
you standardized work is the basis of the
heart of the Toyota Production System (TPS),
which is continuous improvement.

01
Under this philosophy, standardized work Manufacturers who look to new technology as
becomes more than a lifeless document, and a support function for their workforce find
is brought to life by thinking people who greater gains in productivity and quality than
notice deviations from standard and take those who seek technology just for
rapid countermeasures. Thought about in this automation purposes.
way, the benefits of Industry 4.0 technologies
are not to replace people, but to provide Standardized work, a foundational principle
real-time information that amplifies the power for manufacturing efficiency, is one such area
of human creativity to solve problems and that deserves a technological boost in the era
continuously improve. of Industry 4.0 (and beyond). In this
guidebook, we’ll take a look at the evolution of
The role of new technology in standardized work, go deeper into what the
standardized work concept really means, examine how AI and
The attributes that make people a liability in computer vision from Drishti can extend the
assembly environments also make them impact of standardized work and recommend
invaluable. The same qualities that create how manufacturers can layer Drishti on top of
variability in assembly processes also make standardized work to drive continuous
them adaptable to change. In times of improvement, all while keeping workers at the
uncertainty, an organization’s success and center of the equation.
competitive edge is tied to its ability to adapt
to changes.

Here’s where the mindset of automating as

much as possible hurts manufacturers to
some degree: Because we’ve established that
people are in the factory for the foreseeable
future, it makes sense to embrace the role
that people play and invest in technology that
helps them perform to their highest capability.
In short, automate what you can, augment
what you can’t. Instead of looking for
technology that takes people out of the
equation, manufacturers have the opportunity
to instead help people get the job done
better.

In short, automate
what you can,
augment what you
can’t. 02
Contents

1. Standardized work: The evolution of a foundational manufacturing principle

2. Work instructions ≠ standardized work
3. Terms matter: Drishti in a standardized work world
4. Continuous improvement means never achieving “perfect”
5. Continuously improving materials and information flow
6. Using standardized work as a basis for quality improvement
7. Standardized work puts people at the center of the process

Conclusion

About Drishti
Author bios

03
Part 1 Standardized work: The evolution of a
foundational manufacturing principle

Standardized work is the cornerstone of any

lean methodology, so many manufacturers
might not realize that the concept of
standardized work predates TPS by more
than 50 years, and is the product of a
mechanical engineer, not a Toyota original
concept. In 1911, Frederick Winslow Taylor
published “The Principles of Scientific
Management,” which applied the baseline “What is the best way to
rubric of the scientific method to business
do a thing? It is the sum
problems.
of all the good ways we
Instead of randomly applying changes in have discovered up to
order to achieve better outcomes, an the present. It therefore
organization would seek to create controls becomes the standard.
and change only a select few variables to
To decree that today's
check the results. Among other things, Taylor
described in detail how one performs a time
standard shall be
study using a stopwatch and a notebook, a tomorrow's is to exceed
task well known to the modern manufacturer. our power and authority.
Such a decree cannot
Taylor’s concept was fundamentally flawed,
stand. We see all around
though. He assumed that there was one “right
way” to perform a task, and once that way
us yesterday's
was discovered, you trained others on how to standards, but no one
accomplish the prescribed steps and mistakes them for
achieved perfection. The idea of continuous today's. Today's best,
improvement was not on his radar.
which superseded
Using the principles of scientific management,
yesterday's, will be
Henry Ford created his assembly line for the superseded by
Model T and implemented standardized work tomorrow's best.”
on a grand scale, while also understanding
that “the best way” to complete a task - Henry Ford (source)
changed every day as improvements were
implemented.

04
Ford had the vision to increase the volume of
production while maintaining costs that would
allow for the Model T to be affordable to the
average family. Ford made the first modern
assembly line, and therefore the first modern
mass production system. Every contemporary
manufacturing system is deeply rooted in
Taylor’s and Ford’s contributions.
“By design, flow
Taiichi Ohno is considered the father of the systems have an every-
Toyota Production System (TPS), which came thing-works-or-
to fruition in 1948. TPS is an all-encompassing
nothing-works quality
system that envisions the successful
manufacturing organization as a house
which must be
(figure 1) with certain systems as its pillars, all respected and
of which sit atop the foundation of anticipated…it also
standardized work. means that work must
be rigorously
Toyota standardized (by the
Production System work team, not by some
Highest Quality | Lowest Cost | Shortest Lead Times
remote industrial
engineering group) and
Just in Time Jidoka
that employees and
•Continuous •Built-in Quality machines must be
Flow •Error Proofing taught to monitor their
•Takt Time
•Pull System
•Separate
Human &
own work through a
Machine Work series of techniques
commonly called
Standardized Work
poka-yoke, or
Stability mistake-proofing, which
make it impossible for
figure 1: TPS house even one defective part
to be sent ahead to the
Ohno and others built on Ford’s early idea that
standardized work was somewhat fluid; that next step.”
while there might be a “golden way” to
complete a task, the process of continuous - James P. Womack (source)
improvement was ongoing, and therefore that
golden way could change.

Lean manufacturing, an offshoot of TPS, was

first introduced in 1991 by James P. Womack,
Daniel T. Jones and Daniel Roos in their book,
“The Machine That Changed the World.”

05
Whatever the lean methodology in place, New technology and the future of
what they have in common is a foundation of standardized work
standardized work which, though it has Whether you’re a hyper-focused lean
evolved significantly in the past century, is manufacturer striving to higher levels of lean
still rooted in finding the best way to perform maturity, or you use elements of lean and
a task and unifying that process across other improvement methodologies to increase
assembly operations. productivity and quality outside of a specific
discipline, new technologies like Drishti are
Standardized work should come important to the next evolution of
from the worker standardized work in manufacturing.
Frederick Taylor was a mechanical engineer.
He didn’t develop the concept of The tools to understand what’s happening on
standardized work from an office, or as he a manual assembly line, and to determine
was designing product specifications. It came standardized work adherence, have not been
from him observing workers on the floor. As updated for the digital age and are not
standardized work has evolved, and TPS/lean scalable. The data sets are manually gathered
manufacturing rose in popularity, that same via stopwatch and notebook, and are limited,
concept is front and center: standardized biased and based on human cognitive
work should come from the worker. limitations of observation. Simply put, it is
inefficient and ineffective to have an engineer
TPS states that the people doing the work observing every action in a workstation.
know the job the best. They can help guide
the development of standardized work with How can we more effectively measure the
the understanding that there are multiple work being done on the floor to best
ways to complete a task. There is unlikely to understand whether standardized work has
be a static “one best way,” and in fact, the been followed, sustained and improved upon?
point of standardized work is to drive
continuous improvement, making the process Drishti seeks to solve a 100-year-old problem
better over time. by automating the data collection on
assembly lines and presenting that data in
Standardized work alone does not drive insightful, actionable reports and dashboards.
continuous improvement. It sets the The rest of this guide will focus on the
foundation for improvement by establishing a development of this foundational concept
baseline of cycle and action durations and through Industry 4.0 and beyond.
sequences, and ensures consistency across
stations and line associates. But unless
someone is watching, tracking and deriving
insights from that work, nothing changes. The
key is to gather data on the work being
performed and use that data to continually
improve on the assembly process.

Dalia Peña
Customer Success

06
Part 2 Work instructions ≠ standardized work

The Lean Enterprise Institute defines Standardized work is different from work
standardized work in the following way: instructions, which outline how work should
be done, often at a lesser level of detail than
standardized work, and don’t address by
Establishing precise procedures for whom and when the work should be
each operator’s work in a production completed. Even if an organization has a
process, based on three elements: written set of work instructions for every
process and the product comes out within
1. Takt time, which is the rate at specification as a result of those instructions,
which products must be made in a it does not mean that standardized work has
process to meet customer demand. been established.

2. The precise work sequence in There are some cases where standardized
which an operator performs tasks work seems unnecessary: during the
within takt time. prototyping stage, for example, or if current
operations are meeting customer needs
3. The standard inventory, including without standardized work in place. That said,
units in machines, required to keep there are many forcing event that could bring
the process operating smoothly. the need for standardized work into focus,
including but not limited to:
- Lean Enterprise Institute (source)
• Retirements or turnover that result in the
loss of tribal process knowledge

Standardized work provides a system by • Revised customer contracts that demand

which steps in a process are recorded and higher volume or lower defect rates
trained so that the work is done repeatedly, • Moving from prototype to production phase
consistently every time, with the actions
• Identifying a bottleneck station that needs
being done the same way, regardless of the
more granular direction for workers
worker or shift. It is implemented throughout
the manufacturing operation in a deliberate
Work instructions are certainly a part of
manner in the form of a cyclical process:
standardized work packages, but the two are
not equivalent. Proper standardized work aids
• Identifying the process to be standardized
in the removal of information taken for
• Implementing the standardization granted (such as materials placement or
• Tracking and measuring the adherence broadly defined steps) and keys in on
important details.
• Determining whether an improvement is
seen
• Ensuring the improvement can be and is
sustained
• Beginning the process anew

07
Establishing standardized work from Remember, the people who design the
work instructions process are often not the ones who have to
To start, select the process that needs sit and do the process thousands of times a
standardized work defined. Gather the day; it’s a core Drishti tenet that workers often
materials available and go to the place develop better assembly processes. It is
(genba) where the work is done. Some critical that the process is seen as naturally as
processes will have well-defined work possible to determine the validity of current
instructions; others will not. Take the material work instructions (or even as a method of
and watch the workers do the steps. Follow developing them in the first place).
along and determine whether or not they are
doing the same actions. Note the differences: Drishti: A real-time standardized
work audit
• Does one worker use an overhand grip to Once standardized work has been developed
tighten a nut while others use underhand? for a process, review it and confirm its validity
through observation. Drishti is helpful here, as
• Does one worker have his tools in a different
it can automatically detect standardized work
place than the others on the workbench?
deviations and flag cycle and action
• Does one worker adjust the lighting to be anomalies for further scrutiny. It also allows
brighter than other stations? engineers to observe the process from afar,
• Is there a preferential use of tooling that eliminating any performance adjustment on
varies among workers? the part of the worker and removing bias from
the process. It is important to establish
The purpose at this stage is not to establish periodic reviews to ensure that the process is
right or wrong, but simply to note the being followed and still valid.
differences between what the work
instructions say and what is actually being
done. Note this may take several passes to
get correct. It’s important to ask questions of
the worker and establish the why.

08
 figure 2: Drishti flags standardized work deviations and categorizes them as errors or warnings

It is important to approach each process the

same way so that standardized work
documentation remains consistent throughout
the organization. Properly implementing
standardized work applies the fundamentals
of scientific management by establishing
controls and reducing variability in a process.

Diana Fierro
Customer Success

09
Part 3 Terms matter: Drishti in a standardized
work world

When considering standardized work, and

working with a technology provider, it is
important to align on operational terms.

Drishti uses a specific set of terms that align

with standardized work when implementing
its own products on a workstation; many of
these are standard to manufacturers, though
they are not by any means universal. Before
we move on, let’s take a moment to align on
terminology.

Takt time:
Takt time is the rate at which the product
needs to be produced in order to satisfy the
customer demand.

Note: It is not uncommon for organizations to

view each workstation in an assembly line or
each major process in the build of a product
as the basis for takt time. Drishti views takt
time as the target time within which each
individual step has to be completed in order
to meet customer demand.

figure 3: Drishti's Line Balance chart helps manufacturers balance the line for higher productivity
10
 figure 4: Drishti's Line Variability chart shows short and long cycles, and is backed by video

Cycle time: Why does cycle time matter?

In standardized work, each major process is Significantly long or short cycles that deviate
broken down into a number of steps. There from a normal range are indicators that more
may be several job breakdowns in a process. attention should be paid to that cycle. Did the
For the purpose of standardized work, the worker finish the cycle 10 seconds early
total time for all of the major steps in a job because she forgot to complete a step? If so,
breakdown is referred to as station cycle quality issues could be a concern with that
time. Each individual step may have its own unit. Did the cycle take 15 seconds longer
time, as well (called action time, below). than usual because the worker was waiting
on materials? If so, is there a materials flow
Production cycle time issue that could be remedied further
refers to the time it takes to produce a unit at upstream?
the end of the line, which can be measured
by the MES scanning a unit going out to Because Drishti provides video footage of
packaging. Drishti has the capability to show every single cycle, clicking on long or short
both production cycle time and station cycle cycles can quickly show manufacturers why a
time. cycle was deviant, and whether more
attention to that unit, worker or station is
Standardized work focuses on cycle time for warranted.
each process which, ideally, is a summation of
the time in each of the steps plus any waiting
times or movement time it takes for a product
to enter and exit a workstation.

11
 figure 5: Drishti's Yamazumi chart shows how much time each step takes by station

Action time: Why terms matter

Within every cycle is a set of individual Drishti was built on the foundation of
actions that, when performed in sequence, standardized work, and uses these terms
make up standardized work for that station. within its Portal. Drishti’s solution provides the
Each action should take a certain number of constant vigilance and measurement
seconds or minutes to perform. capability of an efficiency engineer and can
view and measure the cycle time of each step
Action sequence: as well as the overall cycle time of a process,
The order of actions matters! When in realtime and report on any deviations from
standardized work is properly defined, it lays the expected standardized work performed.
out not just what should be done, but also in What’s more, the video is available for
what sequence. It may not seem to matter reference to troubleshoot and identify the
whether a worker starts all four screws, then root cause should an issue repeat itself.
screws them all in at once with the pneumatic
screwdriver, or screws each complete piece Whether or not an organization decides to
individually. But if standardized work is implement Drishti on their line, they need to
written properly, the thinking is that one of be familiar with the terms in order to
those two options is more efficient than the effectively systemize standardized work,
other, and therefore the preferred method. particularly the measurement terms. With the
understanding of takt time, cycle time, action
Why do action time and sequence time and action sequence, standardized work
matter? systems and documentation become
Similar to cycle time, long or short actions are consistent and meaningful.
indicative of potential further attention
needed. A short action could have been done
incompletely, while a long action may mean
the worker is poorly equipped with the proper
tools to complete the task, for example. Tasks
performed out of sequence can be indicators Aprameya
of quality or efficiency issues. Manjunath 12
Customer Success
Part 4 Continuous improvement means never
achieving “perfect”

Standardized work is not the ultimate goal; it If an organization undergoes the process of
is the means of creating a process by which implementing a system of standardized work
continuous improvement can happen on the properly, it has taken a huge leap forward in
line. There is always some variability in a continuous improvement. By implementing
process no matter how mature an standardized work, it has removed massive
organization may be. Even with standardized amounts of variability at the line level and
work processes universally established and created a foundation for the rest of its
measured, the way to see benefits is to processes, as well as a baseline from which to
continually audit, review and refine processes. improve. Standardized work then feeds into
the other systems, including preventive
maintenance, supply chain management,
design and development and eventually even
sales processes. It is the foundation of proper
continuous improvement (kaizen) events.

"Today’s
standardization…is the
necessary foundation
on which tomorrow’s
improvements will be
based. If you think
“standardization” as the
best you know today,
but which is to be
improved tomorrow –
you get somewhere.
But if you think of
standards as confining,
then progress stops."

- Henry Ford (source)

13
 figure 6: Drishti Action Detection reports if a step is skipped or performed out of sequence

Takt time and out-of-sequence Did that improvement work?

If cycle times are consistently not to takt time, Once a manufacturer studies the data and
it’s an indicator that something is amiss on the insights into operations that Drishti provides,
line. A deviant cycle can indicate any number he or she can implement changes on the line.
of issues: bottlenecks, a worker rushing, a Drishti then performs as a continuous
mistake being made, stress on the line, real-time auditor of the improvement in
ergonomic issues, and so on. To identify the question. Did the change in operations
problem, the engineer or supervisor needs to actually result in the desired improvement? Is
get a handle on the current condition. it being sustained over time and across
shifts? By studying Drishti data, the
Technology like Drishti can automatically alert manufacturer can quickly tell whether deviant
manufacturers — either through the worker in cycles and steps have been reduced or
real time, or to the supervisor or manager via eliminated and if there are other
regular reports — that cycle times are improvements that could be implemented to
deviating from the desired norm and therefore further tighten the performance of the line.
need attention. Drishti also records and stores
videos that clearly illuminate any standardized
work deviations.

Similarly, steps performed out of sequence, or

steps with abnormally long or short durations,
can be instantly flagged and reviewed, or
summarized through a report. Ideally, the
manufacturer would use Drishti to narrow
down the potential cause of any long or short
cycles, then go to the line and verify.

14
Performing continuous improvement With Drishti, kaizens are far less disruptive to
(kaizen) events the organization. Drishti automatically gathers
To make more holistic improvements across data before and after the event, removing
the line, many manufacturers hold bias and providing far more data points than
periodic continuous improvement events, manually gathering would ever allow.
often referred to as kaizens in the TPS Meanwhile, the kaizen team can focus on the
lexicon. Without technology like Drishti on the high-value portion of the event: discovering
line, an event of this nature relies on opportunities and implementing
snapshots of data from the period leading up improvements.
to the event (again, this data is manually
gathered, so tends to be biased, incomplete Once the solution is
and time-consuming to gather). implemented and standardized work is
updated, Drishti continues to provide
The variability on the line that can be identi- meaningful measurements that will verify
fied in a condensed time period is limited, and whether or not the solution worked. Drishti
the kaizen event itself takes at least a week, dovetails perfectly with standardized work
with several weeks of data gathering and and kaizen events to make continuous
analysis before and after. Famously, improvement seamless for an organization.
production ratios consistently drop after each
process improvement as the team learns new Simply put, when paired with standardized
standardized work instructions. work, Drishti is a powerful tool that provides
huge volumes of data and valuable insights
that enhance the thinking of everyone in the
manufacturing operation.

“Without standardized
work, there is no kaizen.”
- Taiichi Ohno (source)

15
 figure 7: Drishti daily reports help you solve problems with targeted data

Simply put, when paired with standardized

work, Drishti is a powerful tool that provides 1.Identify a
problem or
huge volumes of data and valuable insights opportunity
that enhance the thinking of everyone in the
manufacturing operation. 6.
Standardize 2.Analyze
the the
solution Process

5.Study the 3.Develop

results & an optimal
adjust solution

4.Implement
the
solution
Aldo Corona
Customer Success
(figure 8: Drishti enhances the kaizen process,
particularly in the blue shaded steps.)

16
Part 5 Continuously improving materials and
information flow

In systemizing standardized work, we are By many definitions, materials include

ensuring that the processes defined within an environmental factors, as well. Lighting, fans
organization, particularly a production line, are blowing nearby, shadows from different parts
being done in a predictable and consistent of the equipment, seating and more can all be
way. This has a lot to do with the human significant factors in work variation.
workers performing the work, making sure
that they are doing the individual steps and The case to consider materials flow
actions of a job repeatedly in the same From a business perspective, creating a
manner. Drishti specializes in action proper flow of materials creates a happy
recognition, confirming that these actions are supply chain. Sporadic change is the enemy
stable, and reporting problems in real time. of efficiency. Knowing the location and
This is far more effective than waiting to see a working condition of all your tools is an
bad KPI in the team huddle the next day essential part of the manufacturing flow. This
before the shift starts, or worse, from poor is especially true as manufacturers take on
performance numbers at the end of the various regulatory requirements and need to
period. know the state, location and use history of
their assets. It also makes it easier to train
Though workers are the primary consideration individuals with this level of standardization.
in standardized work, there are other There is less ambiguity when there is "a place
elements to keep in mind. What often gets for everything" and that which is not
overlooked is the flow of both materials and necessary is not present.
information.
Where, when and how of material
Materials flow Organizations implementing standardized
Materials flow refers to several things: first, work may need to rethink what “standard”
the actual items being assembled, the raw really means. The detailed work instructions
materials of sub-assemblies of a product. and job instruction breakdown sheets need to
Arguably, this is the most important facet, as be established, but that’s just one component
without these materials, the bill of materials of standardization. Material needs to show up
(BOM) in a finished product cannot be at the right time, every time, at the same rate
complete. In order for jobs to be built on time, and to the same place in order for a cycle to
the material for the job needs to be available be truly defined. The line associate needs to
on time. Second, material refers to the tools take the material to a known specific place in
that are being used to complete the job, a workstation and use the same tools pulled
including equipment that is being used to do from the same place, each time, in the same
the job build. Equipment can be the conveyor manner. These considerations make material
belt or even the size of a table for a given availability a key component of standardized
workspace. Are there duplicate workstations work.
with different styles of tooling available? For
example, a hammer: Is it the same size and
shape at every workstation?

17
Information flow By undergoing this harmonization, we are
Information flow is a model of information further systemizing the organization and
dissemination within an organization. It views making a rubric by which all systems and
an overall system of processes as well as processes can be developed afterward. By
each step, and asks the most basic of defining the flow of information at a high level,
questions: process design removes assumptions, bias
and tribal knowledge that could prevent a
• What are the physical requirements for this process from being effective. If too much
step? information is taken for granted, the
• Does the process have any dependencies organization is at a high risk of process failure.
(e.g. A must occur before B)? While people are the most important asset of
• How do we know what we know for a given an organization, it is very important that
step of a process (is there documentation)? systems be in place so that no one person
• Who maintains the documentation and how holds all of the information required to make a
are changes made? system function.
• What happens before the process starts?
• What happens after the process ends, and Many organizations have fallen prey to this
how does the responsible party know what mistake, believing that they have a solid
to do? process only to be missing steps when an
• Who is responsible for what, and where is employee goes on vacation or leaves. This is
this information? an indicator that too much information was
• How is this information shared across taken for granted, and that the issue of
individuals, lines, shifts and plants? information flow needs to be addressed.

If at any step in the process we do not know Drishti doesn’t take information for
the requirements, responsibilities and granted
information/resource flow, the model has Drishti aids information flow by delivering key
broken down. Not everyone in the insights in real time without bias, providing an
organization needs to know everything about opportunity for workers, managers and
every process, but all people involved in a engineers to use the results of huge samples
given process need to know their part of it. In of data to focus their improvement activities.
properly addressing information flow when For example, Drishti will highlight where the
designing business processes, we are able to material is placed differently from station to
harmonize the flow of information and ensure station and show immediate variability on the
that there is no hangup. A person involved in line. Toyota’s famous andon system is
a process has an if/then set of options in their centered on workers pulling a cord when they
head for each eventuality, which fits into the see a deviation from standard. Think of Drishti
large overall flow of materials and information. as an electronic, automatic andon that
consistently spots and reports deviations
from standard.

18
Drishti’s collaboration tools and workflows That said, Drishti (or any manufacturing
infuse objective video into active problem technology) will only have maximum
solving. Line supervisors and team leaders effectiveness in an organization where the
use Drishti to document good and bad flow of information is clear and the systems
practices and assign tasks in a collaborative are in place to use the provided insights. Any
interface. Industrial engineers define and solution, no matter how advanced, cannot
prioritize improvement projects in the Portal, make the changes happen on its own.
while also assigning and managing teams. Therefore, new technology is only as good as
Quality engineers collect inputs on quality the organization’s ability to use and
claims within the Portal. The use of a “single disseminate the appropriate information to
source of truth” makes information the right people in a timely manner, and effect
transparent and accessible to all necessary the changes indicated. Thus information flow
parties. is essential.

Primo Garcia
Customer Success

19
Part 6 Using standardized work as a basis for
quality improvement

Standardized work is an important driver of Ergonomics

increased quality within an organization. If With processes standardized, it is possible to
standardized work is not established, efforts find the most streamlined material placement
like root cause analysis take longer and are in order to be most efficient, comfortable and
often resolved less decisively than when healthy for the workers. Identifying potential
standardized work is in place. ergonomic concerns and setting up stations
to be comfortable and efficient for workers is
Standardized work ensures that each process exponentially easier when the process is
is done the same way, every time, by every standardized and Drishti can flag deviations
worker. Because the basis of standardized for further examination.
work is that it represents the current best way
to perform a task, it stands to reason that
deviations from that process are more likely to
result in quality issues.

Uncovering ergonomic concerns

Drishti was able to help its customer,

HELLA, a tier one automotive supplier,
in understanding where its true
bottlenecks were and improving
health and safety for workers. The
HELLA team believed it knew which
station was the bottleneck station,
but upon installing Drishti cameras,
With live streamed and recorded video, Drishti they learned that the real line
can help manufacturers streamline a root slowdowns were occurring further
cause analysis process. If a defective unit is upstream.
discovered, they can use Drishti to pull up
video footage of that unit’s entire assembly At one station, the
lifecycle, watching the steps as they are configuration required workers to lift a
performed and noting any abnormalities such part from the left side of the station
as long or short cycles or steps, or steps with his right hand, requiring a
performed out of sequence. Because Drishti repetitive twisting motion for each
can pinpoint those potential problem areas, assembly. What wasn’t clear in
the quality engineer and team can focus on day-to-day operations became
whether a defect was introduced and take obvious when collecting data over
immediate steps to rectify the problem while hundreds of cycles. HELLA both
simultaneously putting preventive measures in increased throughput and improved
place. worker conditions, with a return on
investment in under six months.

20
 Show and tell: Video-based training
In training a task, it’s important to both tell
and show the process. Telling alone is not
enough to gain all of the special knowledge in
a full sequence of steps. Ideally, the process
will be shown, explained, able to be read and
also (in a full implementation) re-trained on
demand using a solution like Drishti, in which
the whole sequence of steps is available via
video and broken down accordingly.

Drishti also allows supervisors to tag videos

that demonstrate good and bad practice,
making it easy for them to recall a library of
examples to show the worker.

Future verification
After training is complete, it is important that
Training best practices the trainer checks back in with the trainee at
When implementing a process change or periodic intervals to ensure that the process
setting a baseline for new standardized work, is being done correctly. Using Drishti, trainers
the organization must commit to training the can spot-train on the line using video footage
standardized work system to staff and of the worker, similar to when an NFL
planning refresher training at regular intervals. quarterback watches the playback after an
interception or touchdown to learn what went
While implementing standardized work, it is wrong or right.
important that there is a joint effort between
the process team and the workers. Anyone If a picture is worth a thousand words, how
who is a stakeholder in the process should much is AI-powered video worth?
have an opportunity to give feedback. Too
often, organizations rely heavily on the
feedback of those not directly involved in the
process. Remember, the worker should lead
the standardized work process. By
incorporating feedback from those actually
doing the work, the process is better from the
start.

The workers hold key insights for continuous

improvement, since they are the ones
carrying out the process.

Sandeep
Abraham
Customer Success

21
Part 7 Standardized work puts people at the
center of the process
The best standardized work instructions and Drishti is built on standardized work,
systems are led by, and get feedback from, too
the people doing the work. Ultimately, the Drishti has built a new AI-powered way to
worker determines whether or not a digitally conduct constant time and motion
standardized work process is effective. studies, create massive volumes of data from
the assembly process and use that
information to drive actionable insights
toward continuous improvement. Here’s how it
works:

1. Drishti installs cameras on desired

workstations and streams video of your
“Every team member assembly processes. By the way, you can
has the responsibility to access this live video anywhere you have
an internet connection.
stop the line every time
they see something that 2. Drishti’s proprietary neural networks work
is out of standard. in conjunction with our team of engineers
That's how we put the to annotate and recognize key points in
responsibility for quality the process (as in the standardized work
sequence of steps).
in the hands of our team
members.” 3. Drishti data is used to automatically
monitor the process in real time, or after
- Jeffrey K. Liker (source) the fact if preferred. Our system can flag
long and short cycles, and even notify the
worker of step deviations in real time. We
refer to this as action recognition (it’s like
object recognition, but uses video streams
instead of still images).

In fact, one of the most important steps a 4. Drishti automatically shares insights
manufacturer can take to improve its about cycles and individual actions with
processes is to ensure that standardized work key stakeholders, including at the line,
is properly implemented in a systematic team, plant and executive level. These
manner. Standardized work is a foundation on teams use this information to quickly
which the rest of an organization's processes identify improvement areas, increase
can be built, and is the baseline required for productivity, reduce defect rates, conduct
effective problem solving. Every organization training and more.
will have problems, no matter how mature,
and it is the systems in place that will
determine how difficult it is to identify and
solve these problems.
22
Standardized work remains a key part of any Drishti customers have seen massive results,
lean manufacturing methodology. Even with including:
increased automation, 72% of all assembly • 50% lower defect rates on lines with very
tasks are still done by hand. The next wave of low escape defect rates (~0.001%), in just
technology needs to empower workers rather three months
than attempt to replace them. This is what • 15% throughput improvement on already
Industry 4.0 forgets. optimized lines
• 15% lower scrap rates
While humans are the most important asset • 50% faster training times for new workers
for manufacturers because they are • Deterministic root cause analysis in 20
adaptable to change, they create variability. minutes
That’s why technology like Drishti is so key: it
reduces that variability while augmenting Ultimately, whether in areas of productivity,
human workers’ best assets, like cognition quality or training, Drishti promotes a new
and reason. version of continuous improvement that
would not be possible without AI and
Drishti does the equivalent of three tedious computer vision technology. In this way,
tasks rolled into one: Drishti is taking the benefits of standardized
work into Industry 4.0.
• Rather than requiring hours from industrial
engineers, Drishti can continuously “stand
on the line with a clipboard,” recording cycle
times. Engineers can focus on deducing the
far richer and more voluminous data and
recommended insights and making
improvements.

• Rather than spending days, weeks or

months conducting retroactive research into
the cause of a defect, Drishti can help
quality engineers uncover the cause of an
issue in minutes, so quality teams can focus
on implementing fixes and putting
preventive measures in place.

• Drishti provides video evidence of best

practices, allowing faster worker training
and onboarding, as well as on-the-spot
training as needed.

Praveen Kumar
Customer Success

23
About Drishti
Drishti’s AI-powered video analytics technology augments human workers and provides visibility
and insights that transform the pace and impact of manual assembly line improvement. Manufactur-
ers use Drishti to anchor true digital transformation, driving sweeping improvements in quality costs,
efficiency gains & time-to-proficiency for line associate training. And line associates rely on Drishti
to be more consistent and efficient, becoming even more valuable on the factory floor. Drishti has
been recognized as a World Economic Forum Technology Pioneer and a Forbes AI 50 company,
among other recognitions. For more information, visit drishti.com.

About the authors

The authors of this guide consist of the following members of Drishti’s customer success team:

Dalia Peña has more than 12 years of experience in Primo Garcia has more than 15 years ofexperience.
lean manufacturing, primarily in the automotive and He has served as a manufacturing engineering
electronics industries. Prior to Drishti, she worked manager and owned multiple lean initiatives. He
for BRP and Flextronics. Dalia has a bachelors in holds a Bachelor’s Degree in Production
engineering physics from the Universidad Technology from Western Michigan University and
Autonoma de Ciudad Juárez. She resides in a Master’s in Leadership from Central Michigan
California, USA. University. He resides in Michigan, USA.

Diana Fierro is a lean coordinator with a Sandeep Abraham is a talented project manager
demonstrated history of working in the electrical and an experienced practitioner of lean
and electronic manufacturing industry. Diana has a methodologies. He holds a Bachelor’s Degree in
Bachelor’s Degree in Computer Systems Electrical and Electronics Engineering from
Engineering from Instituto Tecnológico Superior de Visvesvaraya Technological University. He resides
Zapopan. Diana is located in Jalisco, Mexico. in Karnataka, India.

Aprameya Manjunath is a business analyst with Praveen Kumar is an experienced practitioner of

experience in engineering and renewable energy. deep learning technologies. He holds a Bachelor’s
He has completed internship positions in in Mechanical Engineering from the Sri Krishna
automotive, aerospace and renewable energy College of Engineering and Technology, a Master
systems. Aprameya holds a B.E. in Mechanical of Science (M.Sc) in Computational Mechanics
Engineering from Dayananda Sagar College of from Swansea University and an MBA from the
Engineering and an M.S. in Renewable Energy from Indian Institute of Management Ahmedabad.
Loughborough University. Aprameya resides in Praveen resides in Karnataka, India.
Karnataka, India.

Aldo Corona is a customer success manager with

many years of experience in industrial engineering
and project management. Aldo holds a Bachelor’s
of Engineering in industrial Engineering from the
Universidad de Guadalajara. He resides in Jalisco,
Mexico.

24