ORLANDO, Fla. — If there were a contest to find the buzzword that sums up the Gartner Supply Chain Symposium, the winner would be clear: AI.
Companies boasted about their artificial intelligence bona fides on the floor of the recently concluded show. The flood of emails sent to journalists covering the event, offering up interviews with executives, almost all touted the AI credentials of the companies being pitched. And with international media focused on the capabilities of AI-driven ChatGPT in recent months, supply chain software providers that didn’t boast in Orlando about how they use AI to differentiate themselves would have seemed woefully behind the times.
“I would say there are about 1,000 percent more mentions of AI this year than last year,” Keith Hartley, the relatively new CEO at LevaData, said of the conference.
Of course, AI within supply chain software is not new. But what is new are the capabilities of ChatGPT and how that generative AI can impact the huge network of supply chain software providers. In a recent paper on the subject, the consulting firm KPMG described generative AI as “an emerging form of AI that can create original articles, essays, images, music, and yes, code, by building on patterns it finds in existing text, audio files, images and software.”
What many people at a meeting like Gartner acknowledge in private is that the rows and rows of exhibitors selling software solutions for the supply chain all have products that for some overwhelming majority of their capabilities — is it 90%? It is more? — all do the same thing. There is a basic foundation of functions that needs to be performed in any sort of supply chain technology.
But it is those last percentage points that distinguish one solution from another and can be the difference between success and failure. And with generative AI on top of earlier AI capabilities, it’s clear that this is one of the areas where the battle will be fought.
IBM demonstrates its Watson solution
One of the more heavily attended events was a presentation by Rob Cushman, senior partner, worldwide leader – supply chain transformation at IBM, who demonstrated the company’s still-fledgling — and mostly internal — generative AI capabilities that are targeted to its supply chain business.
Cushman demonstrated a test case in which IBM’s generative AI capabilities were asked to track parts shortages in a supply chain. IBM has had AI capabilities through its Watson natural language processing service for more than 10 years.
His demonstration asked Watson to track the availability of one particular part, and through a series of clicks and written questions, AI would bring up further details on the availability of that part. “What is the stock?” Cushman said, reviewing some of the questions that might be asked. “What’s on order? Do I see gaps in my supply perspective?” The generative AI capabilities in the supply chain provide “an integrated view of what I as a person who owns this part needs to be focused on today as I come online and sit down at my desk.”
That was a point that came up in several discussions about AI: It’s not necessarily a job killer. As Cushman said, the data on that part’s availability that is processed by Watson lets the user “make an informed decision” on managing the supply chain.
Hartley and LevaData used the occasion of the meeting to roll out the company’s new capabilities that also rely on generative AI. LevaData’s offering support materials sourcing; as he noted during an interview with FreightWaves, a mouse on the table in front of him was manufactured with about 200 individual components, and LevaData’s capabilities assist the company’s clients in the supply chain to secure them.
But as he noted, using AI is not new for LevaData. “We’ve been using AI and statistical modeling for a long time, so it’s great that it’s called AI,” he said. “Now it’s generative AI, which we’re adding and moving that way.”
An example of LevaData’s generative AI capabilities appears on the company’s web page with this kind of dialogue, similar to what Cushman showed in his IBM presentation.
Overhaul is another supply chain software that rolled out a generative AI solution at the Gartner conference, this one in partnership with Microsoft. Overhaul’s value proposition is as a provider of visibility, with particular emphasis on finding immediate risks in a company’s supply chain operations.
Microsoft and Overhaul team up
David Warrick, executive vice president of the Enterprise division at Overhaul, said collaboration with Microsoft enables the company to use the software giant’s ChatGPT capabilities to be bolted onto Overhaul’s RiskGPT product, which identifies risks in a supply chain. Warrick said the product is only being used internally so far.
“We have a number of watch officers who are monitoring our platform at all times looking for alert events, like deviations and unscheduled stops,” Warrick said.
Introducing the abilities of generative AI into the process, Warrick said, means that “all of our agents are now working off exactly the same set of rules that are being interpreted the same way.” Where in the past an incident might be addressed differently by the company’s “watch officers,” Warrick said the generative AI capabilities now available in RiskGPT would result in a more systematic approach to dealing with risk issues that the Overhaul service discovers.
Precisely how the new system will play out will take some time to determine; Warrick said it was introduced in Overhaul’s internal systems only last week.
But he added that the capabilities, when opened up to the network of Overhaul users, will not be an add-on service with a separate subscription fee. It will be a standard part of the company’s product offering.
Karin Stevens, chief marketing and product officer at Overhaul, said she sees a role for generative AI in training on compliance issues. Generative AI would “eliminate the subjectivity” in many instances, she said, “and we can take out the decision-making about ‘What do I next?’”
Dawn Andre, chief product officer of JAGGAER, whose platform manages direct and indirect spending, said in an interview that AI and machine learning usage can have numerous applications that aren’t necessarily as visible as something that is driven by generative AI.
“One of the challenges we have today is that the talent is remote, and we are all trying to find a way to bring that expertise into their day-to-day activities, when you don’t have that person sitting in the cube next to you,” Andre said in an interview at the JAGGAER booth on the symposium floor, talking about a use of AI. That sort of expertise can be built into AI-driven applications a company uses in its internal activities, “and it keeps that user and that professional refining their skill set and building their productivity each time they run a process,” Andre said.
Those people remain key even as AI and machine learning grow, Andre said, noting that those processes can “automate a lot of mundane tasks” while staff members can “focus more on being strategic.”
Although incorporating environmental, social and governance (ESG) principles may be morphing into a political football in certain places, it came up several times over the course of the Gartner conference. None of the references suggested that large portions of the logistics industry has any plans on moving away from integrating ESG into decision-making, whether for internal operations or in selecting companies it does business with.
Andre said she sees AI as being able to provide companies with far more detailed ESG information on possible customers and suppliers.
AI gives a company an improved ability to determine if a counterparty meets a company’s criteria, Andre said, because it can utilize data both from open sources and internal survey information.
“Once we have the data, then [AI] can be used to curate information based on criteria that is unique to me, and whatever initiatives are going on in the priorities for your company,” she said. Through that AI-generated information, Andre said, “the system can source the best candidates to participate in a bid process, not just for a product or lane but somebody who shares the same values.”
Still fighting the spreadsheet
If conventional or generative AI seems like such a huge boost to efficiency, shouldn’t rapid adoption be a breeze?
Don’t count on it, came the repeated answer at the symposium.
At Locus, which describes itself as a dispatch management platform servicing the final mile, Chief Revenue Officer Mehul Kapadia said AI is used to meet a broad goal that “any delivery should be done in a better way than the last one in terms of sustainability and cost, so we use AI and machine learning essentially for that.”
AI at Locus uses what Kapadia called “constraints,” which inform a last-mile driver that a delivery can’t be made at a certain time of day for a particular location, or that a certain type of package requires more than one person to make the delivery. “We use our technology to essentially do a lot of the continuous learning and figure out what combination will work,” he said. “It is a decision-making tool rather than just a transport management tool.”
But that’s internal.
Kapadia echoed a theme that came up repeatedly in discussions: With all the technology thrown at the supply chain today, with all the venture capital money, with all the bright people and their aisles of solutions on the exhibition hall floor, with the prospect of generative AI about to ramp up the technological capabilities further, purveyors of supply chain solutions are still fighting the lowly spreadsheet.
“On-the-ground operations will still take some time before we fully get in,” Kapadia conceded. “I don’t want to put a time frame on it. People and change management will be a big challenge.”
More articles by John Kingston
EY Americas leader: 70% of supply chain digital transformation efforts fail
Quiet quit a product of disengaged supply chain workforce
Fifteen states jump into California’s ongoing AB5 legal battle