Three Misconceptions, and One Popular Lie, About Industrial Automation

October 8, 2020 | 6 min read

Amar Hanspal, CEO, Bright Machines

Lately, it appears to have become fashionable to bash robots and automation.  The talk track goes something along the lines of “more hype than reality” or “humans are better” or “even the smartest have failed.” 

For example, this article about Apple, this one about Tesla, or this one about Boeing. If neither the amazing Apple, the great Elon, nor a giant of industry can’t make factory automation work, then what’s an ordinary person to do, except look wistfully at the replicants in Blade Runner? 

To the authors and proponents of this view, I have two words – chill, dude. 

No one is –or should be — claiming that industrial automation has reached some exalted state of perfection.  The examples about the limitations of automation often point out a robot’s inability to handle variability or small sizes or perform the most exacting applications of glue, etc. 

And they are right.  Automation cannot handle every arbitrary situation. What the articles fail to point out, or even consider, is the vast array of cases where industrial automation repeatedly and reliably performs flawlessly. 

The lack of nuance in the discussion reminds me of the strange place that autonomous driving finds itself in today – if a technology can’t handle 100% of the cases, including all the edge cases, then it is “failing to live up to the promise.”  Never mind the fact that the technology is good enough to handle many situations, and airplanes regularly use similar software-based guidance systems for take-off and landing

For the record, humans don’t perform perfectly either, whether driving or doing factory work. Humans fail when they get tired, bored, or distracted.  Industrial automation projects fail when they are too ambitious in their goals or poorly defined and designed. When properly thought through, industrial automation can handle more cases than thought previously possible. This continues to improve every day. 

Here are some of the assertions made about automation. Three of them could be considered fibs. One is just a lie. 

Fib: Companies haven’t been successful with industrial automation.

Truth:  The examples of successful industrial automation far outnumber the failures.

Just like any new technology, some implementations of this technology have been busts. It’s important to understand that these are the exceptions, not the rule. It almost always comes down to setting unrealistic goals and timelines, and not having the right talent or partner to accomplish your goals. I remember a time when enterprise software would get a black eye, back in the early 2000s.  Those days are now long-forgotten – almost every company today deploys some form of enterprise software through a software-as-a-service (SaaS) approach and is wildly successful doing it. 

What changed?  For one, the technology got better. SaaS took care of deployment and configuration challenges.  For another, companies got smarter, and goals got more realistic.

The same is true for automation.  For every story of a failed implementation, there’s ten like this – a company that makes 630,000 control panels for security systems each year but needs to get that to a million units to satisfy customer demand. Using factory automation, that customer ramped production up to 1.1 million and realized a 34% savings at the same time. Or the manufacturer of coffee makers that was getting a yield of only 60% from one assembly line, and used automation to increase that to 98%.

Fib: Automation is an expensive, time-consuming and bespoke process.

Truth: Thanks to advances in technology, it is increasingly fast, economical and flexible.

The world of industrial automation is reasonably mature.  There are hundreds of system integrators who can build you bespoke systems to carry out a single task, or a series of tasks, millions of times.  This is how Coca-Cola gets its bottles filled, and General Motors gets its cars welded.  The problem with such “hard automation” is that it is only economical on high volume, high capital investment projects.  Products that change frequently – like consumer products or network infrastructure devices – are ill-suited for this type of automation.  Fortunately, sensors and computing infrastructure have now evolved to become useful to problems such as this. Computer vision and machine learning can now be applied to make industrial automation based on standard, configurable products rather than bespoke implementations. 

Fib: Automation eliminates jobs.

Truth: Automation creates jobs.

Job loss is the most persistent knock on industrial automation. It’s easy to understand why. When a factory automates part of its production, it often finds it needs fewer workers to make assemblies or fasten screws. That is not the same as needing fewer workers, period. Without upskilling and without understanding the big picture, you could argue that jobs have been eliminated.  In actuality, different jobs requiring new skills are created. Framing the transformations as job losses misses the point entirely.  Are there fewer horse carriage driver jobs today than in 1890?  Are there more limousine driver jobs today?

Take, for example, package sorting – a highly automated activity. The level of automation at FedEx’s distribution center in Kernersville, North Carolina, is already legendary, with the vast majority of packages moving through conveyor belts, scanners, and sorters without human interruptions. But that facility still hires about 100 people a year, even as automation grows. Has warehouse automation eliminated jobs?  Sure. Amazon relies heavily on automation for picking goods as it seeks to offer 24-hour turnaround for more items. But in the last quarter of 2019, it hired more than 96,000 people for a variety of roles supporting that automation, and it continues to grow.

It is always easy to count the jobs eliminated– Oxford Economics predicts there will be 20 million of those by 2030. It’s not so easy to calculate the jobs that will be created—but the World Economic Forum suggests there will be 133 million of those as soon as 2022. We need to prepare to fill those jobs and train workers to move into them.

Lie: Automation is a technology in search of a problem.

Truth: Automation is solving huge problems every day.

The proponents of industrial automation may be guilty of over-indexing on the hype side of things. We have all seen endless videos of synchronized robots, industry 4.0 PowerPoints, and frightening hounds unleashed by Boston Dynamics.  At some level, you may believe that all this feels like nerd Hollywood gone wild more than practical, problem-solving technology.

But you’d be wrong. Without smart farming, your lettuce and tomatoes would be a little less fresh and abundant. Without highly automated warehouses, you can forget about 24-hour service from Amazon. And without industrial automation, the price and availability of products would be significantly constrained.

Ignore the capabilities of automation, and you deprive yourself of the benefits of one of our more powerful, fast-developing capabilities. 

Winston Churchill warned that “Perfection is the enemy of progress,” and he’s still right.  When we look across manufacturing, we see that automation is enabling efficiencies we’re already beginning to take for granted. In the process, it’s opening new opportunities. It’s our job to figure out how to best use this powerful technology so that it continues to be the enabler of jobs, resiliency, and innovation we need and deserve.

To learn more about our capabilities in building the backbone of AI, visit Bright Machines.

Learn more

The Role of Upskilling in an Automated Future

The Role of Upskilling in an Automated Future

Empowering Women: A Spotlight on Bright Machines' Female Engineers
Bright Ideas

Empowering Women: A Spotlight on Bright Machines' Female Engineers

A New Paradigm for the ‘AI Backbone’
Manufacturing Innovation

A New Paradigm for the ‘AI Backbone’