What’s a supply chain? And where does AI come in?

By | 3 minute read | November 29, 2017

IBM’s new Martian friend just landed on Earth and is excited to learn about all the latest developments in human technology. In this Q&A series, IBM experts explain complicated topics to our Martian (and you).

Mars is a pretty barren place, but there’s so much stuff on Earth! Everywhere I look, there’s something to buy. How does all of it get made? And how does it get from one place to the next?

I was curious to find out, so I called Mark S. Yourek, an IBM solutions leader in the global consumer industry. I asked him to tell me about supply chains, and how AI is making them better.

Greetings, Mark. What makes a supply chain successful?

A successful supply chain is one that moves goods at just the right time to just the right place so there’s just the right amount that people are going to want to buy — not more, not less. If they buy less you have inventory left over and you’ll have to figure out what to do with it. If they want to buy more, you’re going to run out and miss sales.

So how do businesses know what people are going to want to buy and where?

That’s not an easy answer. Traditionally, there’s been sloppiness in supply chains because there’s a mismatch between the amount of goods we think customers are going to buy and what they actually buy.

Why’s that?

Supply chains have always generally been limited to forecasting demand and moving product based on past history of demand for similar products at a given location. The problem is that while past history does inform demand to some degree, the future never resembles the past exactly. There are always new and distinct things that are happening that make tomorrow’s demand look different from anything that has happened in the past.

What kind of things do you mean?

There could be weather this year that didn’t happen last year. There could be some kind of event or some kind of trend that didn’t happen last year. All those things can affect customer behavior and demand for different products on a store-by-store level.

Humans can’t see those things coming?

Well, up until recently, supply chains relied mostly on humans to try to keep on top of all those factors, and then make adjustments to product movements and inventories to match. But this has always been a difficult and expensive way to try to address these issues. You simply can’t employ enough people to keep on top of everything. Even if you did, they couldn’t do a consistent job of it.

People have used computers to manage supply chains in the past, right?

Yes, but all IT systems up until now haven’t used AI, which mean the only data they’ve been able to use has been data structured specifically for IT use. New AI-enabled systems don’t need to have all that explained to them. Sure, they need to be trained on how to interpret data, but AI-enabled systems can tap into unstructured data and figure out what it means in context and figure out what’s significant to these processes at hand.

So AI could help supply chain managers do a better job of staying on top of this stuff?

That’s right. AI can keep on top of unstructured information — hyper-local weather patterns, local events, traffic patterns, social media buzz — that are indicators of customer behavior and demand. AI can also do a great job of keeping on top of the actual operations of the supply chain — truck and delivery movements, maintenance and breakdowns, and issues that may occur in transit or at distribution centers.

What kind of issues?

Disruptions happen all the time. Weather is one. Or you might have a strike at the dock at a port that stops goods from arriving. AI-enabled supply chains can tap into all of those things, interpret them, and advise you on how to adjust to keep your supply chain running as efficiently as possible.

I like things that run smoothly.

Humans and Martians have a lot in common then!