✕

Whether it’s from reshoring operations or trying to find new workers for a factory – manufacturers are being pushed toward automation. In the United States alone 67% of manufacturers said they were unable to attract and retain talent. And, according to the U.S. Chamber of Commerce, as of January 2024, a gap persists, with 622,000 total manufacturing job openings yet to be filled.

Globally, things aren’t looking much better. According to McKinsey, job vacancies per unemployed person increased by more than four times on average between 2010 and 2023. Even though the global labor shortage has eased since the COVID-19 pandemic, the gap in skilled labor is growing. This has led to 75% of European employers not being able to find workers equipped with the right skills in 2023.

With governments around the world now introducing tariffs, this could also impact reshoring efforts, further exacerbating the problem, which is why many are looking at automation as the potential solution.

Just imagine dozens of robots working together autonomously in production lines without a single human in sight. Pictures of fully autonomous factories, also often referred to as dark or lights-out factories, are popping up more and more often as the pressure on manufacturers increases. Still, if you’re familiar with the industry, we’re not even close to this level of autonomy yet.

Technology for this level of autonomy has only recently advanced from theoretical capabilities to becoming practically viable. However, numerous steps remain before manual processes can be successfully transformed into automated ones.

Getting started with automation

When customers first see what our system is capable of, many hand us a wire harness or a bundle of tangled cables and ask if we can automate their cable handling and plugging processes. Naturally, when I show this to my engineers, I can practically see the gears turning as they mentally map out the hundreds of steps required to automate such a complex task.

This is why many people often find automation frustrating — it’s challenging and takes time to implement properly. If your factory relies heavily on manual labor, my advice is to start small and build gradually.

Let’s take, for example, a recent customer we helped in Los Angeles. They mostly relied on manual labor with the exception of two robots. Unfortunately, they were struggling to use the robots efficiently and were stuck with out-of-the-box capabilities.

In their production line, there was a human operator sitting at a station moving metal parts. They would start by checking the weight of the part on a scale, running an electromagnetic test, and hanging them back on the rack.

Now, if you’re at all familiar with industrial automation, you know there are several ways to automate this process. Still, because the parts were different sizes, they knew they needed at least one type of vision system to tackle the problem since the grip on the machine would change each time. The company evaluated our system against several other vision systems and ended up going with ours since the parts were quite reflective and shiny, which was difficult for our competitors to handle.  

Automating this process gave them the confidence to try something even more difficult. Recently, they’ve reached out to us again to automate a second application that’s a bit more complicated. We’ll be picking clear, transparent parts out of a box and then assembling them using our system’s AI. The variances in this application are much more difficult to handle, and if they would have started with this application, the roadblocks might have discouraged them from future automation attempts.

Moving on to advanced industrial automation

Maybe you’re in an industry that’s already highly automated, like the automotive industry. Of course, it would make sense to wonder where you might start in this process since you likely already have automation in your factory.

One of our automotive customers here in California approached us to automate some basic cable plugging for them. They were already comfortable with automation and had the skills and technology to implement it well.

We started on a testing station, which is a standard practice in the manufacturing industry. They were plugging cables into automotive seats. Each line had a human operator, and the goal was to make the task a bit more ergonomic by automating the plugging for them. That meant the workers would just set the cable down on the table and return to the other tasks in the line, so the cable could be in an infinite number of positions, which the robot would need to find.

The task itself isn’t difficult, but the seats and lines continued to get redesigned due to new models, new ways of working, and a variety of other complicated factors. That meant our system got pushed a bit to the wayside since nothing was standardized.

Eventually, the engineers decided that there was one application that stayed the same throughout the changes: Bolt torquing.

They set up our system with a cobot (collaborative robot), so the human operators could work around the system without needing to block off an entire cell. Implementing cobots also is much faster since it provides a back-up plan in case something goes wrong – with the robot or with our system. When downtime is so expensive, the cost of automation is often the risk that it might fail, so having this back-up plan was essential to get buy-in from decision-makers.

This task was also low-hanging fruit, so to say, and since our system doesn’t require any specialized robotic programming languages, they were able to train the software on their own. We only found out about it later when they contacted us to automate bolt torquing on their other lines.

In the meantime, they’ve also managed to stabilize the seat’s design, which means our system will also help automate their cable plugging processes in the near future.

WATCH MATT JONES AT THE ASSEMBLY SHOW

Key factors to consider when automating

If we take both examples, then I would say that there are two key factors to keep in mind when looking to introduce factory automation. The first is to take small steps to build confidence. If we had pushed the first factory, who had little knowledge of industrial automation, to think of creative solutions for our system on their own, it certainly would have felt discouraging. Instead, we pointed them in a clear direction at an easy application.

The second factor is to always have a back-up plan. When downtime can potentially cost millions of dollars, no decision-maker is going to want to block off a cell with technology they haven’t thoroughly tested. Cobots are a clever way around this as they can be pushed or moved away until the technology feels “safer” and more comfortable to use reliably in production.

And, finally, by pairing these two strategies together, you can demonstrate a return on investment in automation technology without forcing anyone to take unnecessary risks. Both companies mentioned earlier didn’t begin with their intended applications but still achieved a return on investment, built confidence in automation technology, and identified clear steps for advancing future automation in their factories.

We’ve moved past the “guinea pig” phase of AI, making adoption much easier. Companies implementing it are no longer early adopters, as AI has become increasingly reliable and robust. This progress isn’t limited to technologies like ours but extends to a wide range of AI applications in manufacturing today.

So, if I could offer any advice, it would be to start with a simple application, try automating it, and make sure you have a backup plan in place in case things go wrong.