How can supply chain leaders supercharge forecasting with the help of AI? Vinayak Ramesh, chief customer officer with Ikigai Labs, describes the current state of the technology.

Supply chain leaders need to adapt to uncertainty in three distinct areas, Ramesh says. One is the ability to incorporate external data, such as tariff levels, interest rates and weather patterns, into the forecast, in addition to historical sales data. A second requirement is the ability to work from limited amounts of data and shorter trends. And a third is being able to forecast for multiple scenarios that might play out.

Artificial intelligence makes it possible to address these complexities, but in a way that augments, not replaces, humans. The output of an AI system is only as good as the data it’s fed, Ramesh notes. And data that’s not accessible by traditional methods, based in part on human experience, needs to be part of that.

That said, “there are only so many data points that a human can process. AI can help get you through all of that work, and present possibilities.”

AI and machine-learning systems get better with experience, but not to the point where they can completely take over the decision-making process. Over time, Ramesh says, AI can take on around 80% of the tasks involved in supply chain planning. That allows humans “to focus on the 20% of cases that have a really high impact.”

It’s important that an AI-driven system be able to explain why it’s coming to a particular conclusion or making a recommendation for action. The result “needs to be interpretable by the end-business user, not just data scientists,” Ramesh says.

Even in a so-called autonomous supply chain, humans will continue to play a role. “For completely new territory,” Ramesh says, “humans have to take over. But they’re freed up to devote more time to that.”