Machines are much better than humans at processing finite data. AI can consider data pools and variables no single person could ever manage alone. However, despite gains made in neural networks as AI and machine learning begins to slowly emulate human thought processes, AI’s inherent shortcomings remain. They include difficulties reacting to unknown situations and being able to take non-binary aspects of the supply chain into account, especially involving situations with no precedent and are missing from a database.

So those who are worried about how long it will be before their jobs are replaced, this concern is probably unwarranted, at least in the near term—although things could change very quickly as AI continues to become much smarter and to learn to think like humans do.

The employment consequences tied to amazing productivity boosts AI can offer have yet to be measured. For example, AI can solve certain planning and other tasks in just a few minutes in the time it would otherwise require a group of supply chain managers working over two weeks to complete. However, machines will still not become competent enough to replace entire job functions in the foreseeable future.

The challenge of decision making

“Computers are really good at making fast decisions and taking into account millions of variables,” Nari Viswanathan, vice president of product management, for River Logic, told EBN. “But they are not good at making strategic decisions with lots of ambiguities and holes in available data.”

Humans will continue to manage high-level strategic planning, such as trying to figure out how to use the supply chain to boost net income and service levels, reducing inventories, Viswanathan said. Human intervention is required to evaluate possible solutions and devising plans based largely on intuition and taking corrective actions as plans are executed, he said.

AI will continue to find more uses in the supply chain for management of specific, yet very high-level tasks. The technology, as described above, is not about to make certain supply chain roles redundant, while organisations will increasingly be able to rely on AI to reduce the number of people and man hours required for very specific tasks.

“I think the potential for AI to help decision makers have better, more complete information and the ability of machine learning to understand from historical data when an error occurs and alert a decision maker to the issue are there. But the jump from catching an error and sending an alert to making the management decisions to rectify the situation is enormous,” Seth Lippincott, an analyst for Nucleus Research said. “Managing a production line where the universe of things that can go wrong is limited and the environment is relatively static is one thing. But the complexities of modern supply chains are far too much for AI to tackle in the near-term.”