Getting to the Truth of Automation
As long as there’s been manufacturing, there’s been a desire to improve it. The cotton gin, steam engine, assembly line, and robot all had the same intent: To make things better, faster, and for less money.
Generative artificial intelligence is just the latest and it comes with the same hopes and fears. Executives feel they need it to compete. Workers worry about being replaced. There’s a truth to it, but technology isn’t a guarantee of anything, good or bad. It depends on how it’s used and if it’s improving a company’s bottom line. That determination and of “how technology shapes the world in which we live, and how it affects regions, firms and workers,” is the focus of Ben Armstrong, executive director of MIT’s Industrial Performance Center.
Automation comes with a lot of confusion and pressure. It also comes with certain assumptions. Armstrong’s work is, in part, challenging those assumptions by asking questions, talking to those who are most affected by technology, and providing some context, with the goal of allowing executives and managers to make smarter decisions on implementing whatever the latest technology may be.
Searching for Positive-Sum Automation There are plenty of surveys concerning how people feel about technology. Typically, they involve talking to who’s in charge. The problem? “We often miss the frontline worker’s perspective,” Armstrong says. To counteract that, he created one, asking thousands of workers, across nine different countries, how they actually felt.
One assumption is that employees prefer a bottom-up approach. Some want that input on design and integration, but there’s also a percentage who cares more about comfort and ergonomics. Armstrong’s work isn’t about identifying the best hire or practice. It’s more of a heads-up to managers.
“We’re not saying that one type of worker is going to be more successful with technology or more valuable,” he says. “It’s instead that each of these types of workers is going to be present in an organization and need to be respected as such.”
Another long-time narrative is that employees worry that technology will take their jobs. It’s not unfounded – a chatbot could well impact a marketer or writer – but it’s more nuanced. Some workers, he found, saw that automation could free them up to learn new skills and advance. He also saw that geography plays a part. In Europe, workers are generally more optimistic about technology; in the United States, less so.
One aspect that can affect that attitude is changing how technology is quantified. Robots are not new; they’re fifty years old, but their success is still being judged on whether they can outperform a person. It’s a zero-sum game, he says, where productivity may go up but the company is locked into an inflexible process.
Armstrong, along with colleague Julie Shah, is trying to figure out if there’s a positive-sum game to be had, where executives, managers, and workers see automation as helping them with their existing work, not as a means to replace it. An Ohio manufacturing company offers an example of investing in internal skills. Workers know how to operate the robots and can shift them to new tasks when needed. The employees become more valuable, but the robots do as well and the company becomes more productive.
And that’s a missing piece. “Good automation” has usually been seen as what benefits the employee, but without helping the company, “Those jobs would not be long for the world,” he says. But when it’s mutual, it builds on itself. As seen in Europe, companies that put automation to work on lower tasks end up hiring more people and the companies become more profitable.
Success Doesn’t Just Happen Technology could take over repetitive tasks, Armstrong says, but basic operations still need to be known. Machines make a certain sound when they’re running well or about to break down. Then they need to be fixed and analyzed so the problem doesn’t recur. Workers with mechanical knowledge and deep experience, not software experts, know how to do that, and it pays for executives and managers to remember this.
BMW already had automated factories in Germany when it started building in the United States in the early 1990s. But, he says, the company started slowly, because in South Carolina, plant employees, along with suppliers, didn’t have robotic skills yet.
Compare that to GM, which in the 1980s built a completed automated factory. “It backfired,” says Armstrong, adding that the takeaway of patience ties into the MIT mantra of “moving slowly so you can move fast.”
But it can be hard to move slowly. Generative AI is just the latest shiny hammer. While it can help, executives have to consider if it benefits their mission. Go back to a car company. A chatbot might respond to four customers rather than two in the same amount of time, a measurable increase, but it might not lead to more sales, raising a red flag as to whether the billion dollars of investment is a good one.
But take Trader Joe’s. Retail “has craved new technology,” Armstrong says, but the grocery store decided to stay low-tech and focus on its core competence, and for that, it makes more per square foot than its competitors. It pays to not get mesmerized because AI will always provide a quick answer. Whether it’s the best one is another matter. “That relies on a lot of collaboration between teams of those who really know the process,” he says.
Rethinking Risk None of this is to downplay automation’s potential. Armstrong sees this is as a crucial time for manufacturing, an industry that has a long-standing approach: If something is successful, don’t change it. That conservative mindset doesn’t make investors flock.
What’s needed is a willingness to try and fail. Companies like Tesla and Rivian have provided a blueprint. “If it doesn’t work, they’re willing to tear it down and start fresh.” It’s a new approach but most likely a necessary one.
For decades, US companies have had ideas, but manufacturing has to be done elsewhere. Armstrong says there’s no reason products like diapers, boats, and electric vehicles can’t be made domestically. It just requires companies, often smaller ones that bring a “new energy”, to break from the norm. The resultant high-growth manufacturing would start a loop: Investors would be attracted. Politicians would support incentives. More companies would experiment.
“I think that would be transformative for the American economy,” Armstrong says.