Rhetoric about robots taking jobs is nothing new: From Oxford University to Elon Musk, the fear of AI as an existential threat abounds. But here’s the more complicated truth: It depends on which sector you work in. Some will be far more impacted than others – and that includes manufacturing.

That’s why small business leaders in manufacturing should care about these questions: “How will AI impact the industry?” and “How can I use it to my business’ advantage?”

How manufacturing leaders use AI today

A forecast by Gartner (available for clients) predicts that AI will eliminate more jobs than it creates through 2019 (mostly in manufacturing). That said, AI-related job creation will cross into positive territory in 2020, reaching two million net jobs in 2025. There’s just one problem: Manufacturing isn’t forecast to benefit from this growth.

Gartner predicts that healthcare, education, and the public sector will see great gains from jobs in AI. Meanwhile, manufacturing will disproportionately contribute to widespread job loss: 938,000 manufacturing jobs will cease to exist by the end of 2019.

How are small business manufacturing leaders preparing for these changes? Gartner (GetApp’s parent company) surveyed 699 small and midsize business (SMB) leaders in the U.S. between April 19 and May 15, 2017. (You can learn more about this survey in our methodology section at the end.)

All respondents were required to have significant influence over business decisions. So, we asked them how they’re using new technologies like AI and plan to use them in the future. Within our sample size, here’s what 175 manufacturing and natural resource leaders said when asked how they’re using AI:

Graph showing use of artificial intelligence in manufacturing by U.S. small and midsize business owners

Source: Gartner

More than six in 10 small business manufacturing and natural resource leaders – 61 percent – currently use AI or plan to use it in the next one to two years. These results align with the group’s three biggest challenges to achieving their business goals:

Graph showing the biggest challenges that manufacturing leaders have to achieving their business goals

Source: Gartner

Nearly one in three leaders in manufacturing and natural resources say that using the right technologies is the biggest challenge to achieving their business goals.

These results imply that some manufacturing leaders are unsure how they can gain the most value from using AI within their businesses. To answer this question, GetApp spoke with one of the world’s leading experts on this subject.

How to prepare for the future of artificial intelligence in manufacturing

In 2016, Stephen Pratt left his job deploying all Watson implementations for IBM Global Business Services. Watson, a question-answering (QA) tool, was one of the first AI tools to earn widespread interest after becoming the champion of Jeopardy in 2011. So, it surprised many that Pratt, who had previously founded the consulting arm of Infosys, would leave this role.

Instead, he started Noodle.ai, an enterprise AI software company. Noodle.ai uses interconnected learning algorithms to manage supply and demand planning. Noodle.ai made headlines when it worked with Big River Steel. Their collaboration produced what Noodle.ai calls the “world’s first smart steel production facility” last year.

Today, Big River Steel uses several machine learning techniques (including sourcing and inventory management) to improve its mills’ performance and profitability. AI in the steel industry isn’t a new thing— neural networks were installed back in the ’90s. But Noodle.ai’s partnership with Big River Steel is a case study of what’s possible today. If steel mills can use AI to fix their outbound supply chains, threats from automation might not be so severe.

Read GetApp’s interview with Pratt to learn:

  • The biggest barriers that businesses face to implement artificial intelligence in manufacturing;
  • How “learning steel mills” can improve products and efficiency materials;
  • How AI can reduce the energy costs of operating a steel mill;
  • Why AI-powered manufacturing mills are inevitable;
  • The difference between AI and robotic process automation (RPA) – and why small business owners shouldn’t fear the former.

A recent Gartner survey on the top technology trends for SMBs showed that 27 percent of SMB manufacturing leaders currently use AI, while an additional 34 percent plan to use AI in the next one to two years.

What’s your response to these numbers in light of Gartner’s forecast that the manufacturing industry will be heavily impacted by AI?

“These numbers sound right. SMBs will need some hand holding from outside entities to get into AI, because they don’t have either the internal talent, the budget, or the technical knowledge in the C-suite. Talent is especially a challenge right now. Data scientists and engineers are in high demand. Unless you can provide a good stream of research opportunities, it’s hard to attract and retain top-tier data science talent.

“Some industries will go faster than others because they’re more data oriented and have already made investments in data that will serve as fuel for AI work. [At the same time], many companies will find that their industry is starting to leave them behind. AI is a competitive advantage they can’t afford to miss, and learning algorithms are continually improving. So, it’s very difficult to overcome competitors’ head starts.

“Wide recognition of our work at Big River Steel has been a wakeup call for many manufacturers. We’ve recently spoken to dozens of metal manufacturers who are very eager to ramp up their AI efforts. Even by the end of this year, I expect we’ll see significant increases in adoption.”

Describe how Noodle.ai started working with Big River Steel. How did this collaboration begin, and what was the AI implementation strategy?

“I first met Dave Stickler, Big River Steel’s CEO, while he was attending a manufacturing conference that I was speaking at. Dave loves to describe Big River Steel as ‘really a technology company that just happens to make steel.’

“He had a vision of a steel mill that ingested massive amounts of data alongside the raw materials for steel production. I knew that Noodle.ai could bring this to life. A learning steel mill could absorb data from production sensors, but also combine it with external signals – market data on raw materials or finished goods – to continually improve product quality and operational efficiency.

“A mill like Big River Steel can use as much electricity as a small city. So, not only was there an opportunity to improve margins with better demand signals or production improvements: We could positively impact the environment as well. Better quality steel that is safely and responsibly manufactured means a more competitive product and an increase in demand and ultimately jobs.”

Give one example of a specific AI technology that Big River Steel implemented and a concrete business outcome attributed to that implementation. (i.e. “Big River Steel lowered overhead by 25 percent within one year of implementing deep neural networks (DNN)”.)

“Energy is one of the primary costs of operating a steel mill, and understanding consumption in detail is critically important for managing cost.

“Noodle.ai has worked with Big River Steel to configure an AI-driven energy usage prediction and shaping application to predict future hour-by-hour energy consumption and enable dynamic demand-shaping based on its daily production schedules.

“As a result, Big River Steel, the world’s first LEED-certified steel production facility, will realize a seven percent reduction in its energy consumption and costs. That’s the equivalent of about 46,000 tons of CO2 per year.”

How can steel mill owners/operators be better prepared to tackle business in the next five years? How will AI impact the overall steel industry?

“We always advise any customer to start by educating themselves. AI is truly transformative, but it’s not magic. There are a lot of companies selling AI software that is not going to significantly impact the bottom line. Educate yourself on what AI can and can’t do, and prioritize use cases that can provide demonstrable impact.

“As AI matures within steel, we expect to see an increase in the impact of business decisions powered by AI. It will become easier to predict outcomes resulting from operational changes. [For example], how would changes in my delivery schedule affect overall energy consumption? Or, how can I hedge on raw material prices to improve my inventory management?

“Eventually, the entire factory will be supported by continuously learning machines. That’s the only way steel mill owners, or any manufacturers, will survive in a globally competitive world. They need to become as efficient as possible in every conceivable way, from order taking to logistics.”

Do you predict job loss in the steel mill industry from AI implementation?

“Imbuing steel mills with AI will have the effect of improving production efficiency and responsiveness to demand. Better quality at lesser cost is generally a recipe for increasing overall demand. [This will create potential] new markets for steel. While demand for steel increases, steel jobs won’t go away.

“People often confuse artificial intelligence with robotic process automation. But enterprise artificial intelligence – the AI that we’re focused on — is about realizing the value hidden in data that would be impractical for humans to utilize by any other means.

“Enterprise artificial intelligence is a means to leverage every byte of internal and external data so that operators can make deeply informed business decisions. Steel operations are already very reliant on technology: These machines are just getting a lot smarter.”

How does AI impact other sectors?

Methodology

Gartner conducted an online survey between April 19 and May 15, 2017. The survey received a total of 699 respondents in the U.S. They were all required to have more than 10 employees and annual revenue less than $100 million USD.

Respondents were also required to have significant influence on business decisions. Within this dataset, 181 respondents worked in the manufacturing and natural sources industry.