Industrial AI: The Technology’s Role Grows in Heavy Industries

Much of the conversation surrounding AI over the past few years has centered on fields like legal, healthcare, marketing and finance. But industrial companies have been using automation to improve efficiency for decades. Now, generative AI is giving these organizations another tool to help them achieve their most critical business goals.

In these sectors, reliability is non-negotiable, as industrial AI often involves decisions that can impact worker safety, production uptime and supply chain reliability, notes Jeffrey Hojlo, research vice president for industrial ecosystems, business networks and manufacturing insights at IDC. “The difference is that these AI use cases are mission-critical, impacting human safety and the ability to serve the end customer,” Hojlo says.

READ MORE: IT infrastructure modernization increases agility and efficiency.

AI Delivers a 20% Productivity Boost

The ITS Logistics AI tools, along with the rest of the company’s workloads, run on Microsoft Azure. One AI solution, a carrier pricing model, automatically recommends carriers and target prices based on performance, cost and reliability. The system uses data from 600,000 annual shipments, including on-time delivery rates, driver safety records and historical pricing data, and it is fully integrated into the company’s existing transportation management system.

“The return on investment on that project alone has been over 700%,” says Weis. “That would cover the entire cost of our AI initiatives. It’s been a grand slam for us.”

ITS Engage enables users to track and manage all facets of their supply chain, encompassing imports, over-the-road transportation, and order fulfillment. Other AI solutions include a strategic sourcing model, which uses highway camera data and third-party feeds to identify optimal carriers for specific freight lanes, and a trailer optimization model, which predicts where the company should position its 6,000 trailers based on seasonal demand and customer orders. It is also launching a generative AI chatbot trained on company information that will help employees to make natural language queries.

Weis says the company is shooting for a productivity improvement of around 20%. “If you look at the size of our company, a 20% improvement would save us millions of dollars, and that would go straight to the bottom line. And our users would be more engaged, because they would feel like they’ve got great tech to do their jobs.”

The company, Weis says, has virtually no technical debt and IT leaders spent two years carefully building out its AI platform rather than rushing solutions to market. “That’s allowing us to launch game-changing features in months, not years, and spend tens of thousands of dollars instead of millions,” Weis says. “Now we’re starting to see the payback, and it is sweet.”

UP NEXT: Businesses can build AI centers of excellence.

How AI Helps Jabil Accelerate Sales Cycles

Jabil, a global contract manufacturing and engineering services company based in St. Petersburg, Fla., has been using automation on factory floors virtually since the company’s founding in 1966, says CIO Chase Christensen.

For most of Jabil’s history, that automation has come in the form of machine learning and computer vision. But over the past few years, Jabil has built several generative AI applications using Amazon Q Business by Amazon Web Services. One of these tools, an intelligent shop floor assistant, is trained on the company’s standard operating procedures, troubleshooting guide and technical documents, and workers can query it for answers when unexpected issues arise.

“The worst thing that can happen to us as a manufacturer is downtime,” Christensen says. “Instead of immediately calling someone when they see a light go off, operators can do the first triage. They can course-correct it themselves. Or, when they call a senior engineer to help, they’ll already have some context. That makes for a faster resolution.”

Another AI tool, a procurement assistant, brings LLM capabilities to Jabil’s existing procurement platform, helping employees to identify opportunities to save on materials. And an AI-powered customer research assistant is focused on helping Jabil to shorten its sales cycles.

“Our sales professionals and business development teams can use the agent to more effectively get to customer and market information faster,” Christensen says. “Instead of taking days and weeks with a couple of analysts, they can now get there in minutes. If you don’t move as quickly as your competitors, then someone else wins out.”

RELATED: Bridge information gaps with CDW’s technology support services.

How To Ensure Companywide AI Buy-In

Before rolling out any AI tools, Jabil established a data and AI council sponsored by the CEO and Christiansen, bringing together senior executives to define policies and compliance frameworks. The goal was to enable innovation while ensuring data privacy across the healthcare, government and other sensitive sectors where Jabil operates. The company also made efforts to communicate the goals of the AI tools across the company, as well as to train employees to use them.

“We had to focus on data literacy, making sure that every individual employee understands what AI is and how to interface with it — and what it can and can’t do,” Christensen says. “One challenge was adoption, just the fear of the unknown. We had to double down on making sure that everyone understood that their data was not being exposed, and that we were managing compliance and caring about data privacy and security.”

Even as Jabil is seeing business benefits from its generative AI solutions, Christensen notes that industrial companies must continue to invest in other AI solutions, such as computer vision, which is critical for automating inspections.

“When it comes to some of our products, even a little microscopic scratch that you wouldn’t see with your eyes does impact the performance of that system,” he notes. “We need to continue to look at computer vision, standard statistical modeling and machine learning, along with generative AI. Together, these technologies are going to create great change for industrial companies over time.”