Are modern-day supply chains really that smart? Heidi Benko, vice president of solutions strategy and marketing with Infor Supply Chain, describes how far we’ve come in creating automated systems that equal — and surpass — human capabilities for managing the supply chain.
SCB: What does the term “intelligent supply chain” mean to you?
Benko: It’s based on the supply chain that’s data-driven in nature. It leverages data and insights, and applies machine learning and artificial intelligence, to drive processes and decisions. We’re at this point now because supply chains have changed so rapidly, but also because we have a lot of new digital technologies available. We’re getting real-time signals from sensors, versus a human interpreting and updating the data. And we’re applying machine learning to identify new patterns and insights, and help users make smarter decisions.
SCB: Data by itself is “dumb.” How does the machine become intelligent, as opposed to simply processing data and spitting it back out again at another place?
Benko: To get to the intelligent supply chain, a digital foundation is required. That means no more emails, phone calls and spreadsheets. The end-to-end process has to be digital.
SCB: Taking people out of the process entirely?
Benko: You want good data, and for the machines to learn, but humans still have to interpret the interactions. The algorithm starts to self-tune, and become more predictive and prescriptive. The goal over time is to run autonomously where it makes sense, for routine everyday tasks.
SCB: So you start with a system that’s descriptive in nature, and eventually move to one that’s prescriptive?
Benko: Yes. Some of our customers are just getting to descriptive. They’re all moving from manual to digital processes in some way. Then we apply machine learning to determine when shipments will arrive and be delivered. That gets us the predictive piece, based on all that data. Then next is to prescribe what to do.
SCB: With machine learning, does that mean we’re talking about a diminishing role for humans and an increasing one for machines?
Benko: Humans can only process so much data. If you look at the massive amount of data that machines can learn from, they’re going to find insights that humans just couldn’t. And that’s going to give information to humans to make better decisions. You’re always going to need the human element to interpret what to do with something. “Autonomous” doesn’t mean no people. It means smarter, more efficient people.
SCB: How important is it to know how the automated system came to a particular conclusion?
Benko: You have to know. A lot of data needs to be tuned initially to get the models to work. Someone has to look at that data and interpret it. But it’s also a question of building trust in the machine. You start to see that what the system predicts consistently outperforms what you’ve manually predicted, so you begin letting the system prescribe actions. You trust that those prescriptions are better than what you can do.
SCB: So where’s it all going ultimately? How can we expect these systems to further evolve?
Benko: It still has a long way to go. The intelligent supply chain should help me deliver to my customers as best as I can — at the best cost, on time, and with the best service — and even identify new products. And it’s not just one piece — it’s the entire value chain.