From cost savings and revenue growth to end-to-end visibility of manufacturing operations and waste reduction, the use of artificial intelligence (AI) tools is vital for manufacturers to stay competitive.
Learning how to identify and deploy AI solutions comes with its own challenges, requiring a smart strategy to invest in the right tools and resources aligned with business goals. Simply put, there is no one-size-fits-all approach.
“Depending on the type of manufacturer that wants to use artificial intelligence, the use case is going to be different,” said Graham Immerman, chief commercial officer of Northampton, Mass.-based MachineMetrics. “We find that, in general, manufacturing is five to seven years behind other industries in terms of adopting new technologies like AI and machine learning (ML). In short, the opportunity exists to use AI but identifying achievable use cases that can create rapid value is a critical first step.”
For some companies, using AI and ML begins and ends with predictive maintenance, leaving the vast potential of their use cases untapped.
“When we are able to pull huge amounts of high-frequency data from a customer’s tooling processes, we can look at it one thousand times closer than the standard level to see potential chips or microfractures that were just not visible before,” Immerman explained. “With this information at our fingertips, we can apply machine learning to this data to identify issues before they happen, which saves our clients time and money.”
In the past, it was easy to predict when a machine might fail or break, but those predictions were often too late. In other words, the break was significant enough that the machine was going to have to be shut down anyway, even if there was insight into minor imperfections. That process now seems quaint thanks to an increase in the capability to pull new data.
With competitors vying for a limited pool of skilled labor, and institutional knowledge at risk of leaving with those heading into retirement, the need to use AI effectively is increasingly important.
“As more companies understand how machine learning can make them smarter, suddenly they realize there can be advantages beyond fixing parts,” Immerman noted. “Part counts and cycle times have always been tracked and logged by humans. But as platforms like MachineMetrics continue to aggregate data, learning tendencies, and practices, the AI can suggest why a machine is down, help with better quoting, adapt scheduling, adjust sales forecasts, all of which address inefficiencies and improve the business without any manual data entry required.”
For many companies, access to data and deep insights is all they need to invest in artificial intelligence, but many companies need direction to understand what level of investment will bring the most value.
In a report by McKinsey & Company, “Capturing the true value of Industry 4.0,” the group breaks down the three types of companies poised to join the AI race.
“Cautious starters” are comprised of companies just beginning to explore a digital transformation journey. They need support when it comes to seeing the big picture, because they want to have a strategy that can be deployed across their entire network.
“Frustrated experimenters” have started dabbling with some success but remain frustrated with a lack of clear understanding about how to achieve or measure ROI.
“Ready-to-scalers” are deploying solutions and technologies but struggle to get full returns at scale. They need to find a way to refocus and capture the full benefits that they know are available, while also responding to shifts in business and customer needs.
From those dipping a toe into AI to companies ready to scale, being able to extract keen insights from data is the backbone of a digital transformation strategy.
Quantiphi, an AI-first digital engineering company with U.S. operations based in Boston, has been at the forefront of empowering manufacturers with access to unified data across disparate locations.
“The cloud has enabled manufacturers worldwide to leverage data-driven AI solutions that reduce manual effort, avoid human error, and increase operational efficiencies,” said Saumil Singh, global manufacturing lead for Quantiphi. “In an industry where efficiency applies to everything from supply chains to production lines and customer delivery, embedding data-led intelligence has become imperative.”
According to Singh, businesses are accumulating data at an unprecedented rate, and harnessing this data accurately has become the core of smart manufacturing.
“Prioritizing efforts to rapidly and securely leverage this mountain of vital manufacturing data is helping businesses build better products, stay competitive, enable AI-led innovation, and be better prepared for the rapidly evolving future.”
The integration of AI has helped manufacturers realize the value of data they already have and how intelligent solutions can help monetize this data to achieve lower operational costs and optimize mission-critical workflows. Quantiphi identified four ways in which manufacturers can scale cutting-edge AI into all facets of the manufacturing ecosystem.
The first is engaging with manufacturers to measure and regulate manufacturing operations. One of the toughest challenges manufacturers face is collecting machine and plant data to see and measure everything across the value chain. Quantiphi works with several Fortune 500 companies to leverage AI and enable them to maximize plant availability, productivity, and quality. Thanks to the recent advances in technology, initiatives that took years in the past can now be achieved within months with the power of AI/ML.
The second way AI can help relates to the supply chain, where AI is being used by organizations globally to overcome challenges around demand forecasting, planning, and logistics by leveraging historical data, economic outlook, geopolitical scenarios, and weather conditions.
A third aspect revolves around optimizing business operations. AI-powered solutions such as Document AI, which automates and validates documents to streamline workflows and reduce guesswork, drastically reduce processing times and help regulate business operations. Data can also be used to augment customer experiences and reduce customer acquisition time.
“Data generated from customer interaction touchpoints can be leveraged for customer profiling, churn prediction, and other use cases that help with revenue enhancement,” Singh explained. “Businesses can further leverage AI and machine learning to enrich the customer experience with capabilities like chatbots and personalized recommendations, just to name a few.”
The fourth, and perhaps the largest, segment where AI is advancing is around sustainability. “Efficiency can be tied to sustainability as more businesses aspire to make this a part of their underlying processes,” Singh added. The efficiencies brought into the manufacturing and supply chain workflows largely impact the energy footprint of a business. “It’s clear that AI is one of the fastest-developing technologies today with the most diverse applications relevant across industries. But the whole point of bringing AI and machine learning into the manufacturing landscape is to facilitate more efficient operations.”
With more legislation passed to move the needle on sustainable business practices, AI is poised to play a larger role than ever in a manufacturer’s sustainability efforts.
“Five years ago, the odds that the sustainability director knew anyone on the shop floor were pretty low,” conceded Rick Oppedisano, CEO, Delta Bravo AI. “That has all changed in the past handful of years as sustainability rises to the forefront and more attention is paid to it, so artificial intelligence has to follow.”
With supply chains still not operating at maximum capacity—and perhaps never getting back to pre-pandemic levels—manufacturers are considering using more recycled materials to meet customer needs because they can be more readily available. The issue, however, is that there can be big differences between raw and recycled materials that are used in the specifications of the manufactured product.
Using raw materials, the manufacturer is working with a controlled variable in that the company knows where it’s sourced and how the material will react to certain processes. With a recycled material, the specifications are more varied. The manufacturer might not know where it comes from or if the quality is lower. The material will still fit into the finished part, but it takes longer to distill how it fits. That’s a process that needs to be buttoned up.
When AI enters into the equation, it can analyze how recycled material interacts with raw materials, giving the manufacturer an atomic-level view that is critical to ensuring everything works together.
“We are seeing a lot of manufacturers use AI and machine learning to adjust processes and specifications in real time to improve the quality of the finished goods that use recycled materials instead of traditional raw materials,” Oppedisano said. “One of our customers, Nucor Steel, weaved AI into their innovative industrial steel process and has seen cost savings and energy reduction that just didn’t exist before, and they have artificial intelligence to thank for it.”
In 2022, the rising cost of materials created pressure on plant margins within the steel industry. Nucor, based in Charlotte, N.C., realized that by reducing steel scrap even by a small percentage, the company could save hundreds of thousands of dollars and increase the plant’s throughput.
But Nucor makes more than 1,000 different products that combine various elements. It needed a way to predict how each mixture would perform and get an accurate recommendation on where to cut the mixture when it would reach optimum quality levels.
By working closely with Delta Bravo, Nucor was able to implement AI and access a graphic display of predicted behavior for each element and product, along with a recommendation of where to cut steel mixtures for optimal quality. This meant the company could blend together raw and recycled materials to produce the same quality of steel that customers have come to expect.
It showed how valuable AI can be when it’s rolled out properly and used for a specific purpose within a manufacturing process.
“We’re saving money, and, by reducing rework, we’re also reducing our energy footprint. It’s the kind of investment that’s perfect for Nucor—it’s innovative, different, and drives measurable savings,” said Dylan Clark, metallurgist at Nucor.
While AI has passed the hype stage, there is work to do to demonstrate its tangible value for manufacturers to invest time, money, and other resources. So where does it go from here?
That’s a question on the minds of every manufacturer that continues to deal with delays and issues caused by a global pandemic, uncertain economy, and unpredictable supply chain.
“Over the last three years, our customers have been exclusively focused on survival, and for good reason, as unprecedented uncertainty has thrown a wrench in plans,” said Jerry Foster, chief technology officer of Plex Systems Inc., a Rockwell Automation company and leader in cloud-delivered smart manufacturing solutions. “As we move further away from global upheaval in the manufacturing space, we think a lot of companies are poised to take advantage of new artificial intelligence offerings.”
One way customers are benefiting from advances in AI is identifying patterns that were not visible before. As AI and ML evolved, comparing data between companies in different industries was not only feasible, but useful.
“When it comes to machine learning, the whole premise behind this solution is to uncover patterns you didn’t see before,” said Foster. “When you add contextual data around something, we can hone in on better information that can be deployed broadly.”
Foster explained that when an algorithm is created using AI, Plex might see trends from one customer that match those of other customers in different areas. These flags alert users to the existence of similarities where they wouldn’t think to look, giving customers insights they never imagined.
“This is the type of smart configuration tool we can test internally with different customers to see if there is something worth pursuing,” said Foster. “When we deem it to work, we roll it out broadly and allow customers to take full advantage.”
The next frontier might be adaptive AI, or AI that writes itself, Foster explained. “Whereas traditional AI can be trained using data to make predictions, the next step is for AI to learn over time to recognize patterns on its own and share them with the user.”
While this might seem like a far-fetched idea, it is within reach, Foster asserted, and will allow companies to benefit from being able to discern key learnings without the help of specialized data scientists.
At a time when manufacturers endure labor shortages and myriad disruptions while trying to do right by customers, AI and machine learning are tools to help, and successful companies will look beyond the hype and understand how to get the most out of them.
The next step in artificial intelligence will be giving the technology a longer leash to let it do what it’s trained to do, which is to gain insight and help humans make better decisions.
Connect With Us
