Watch: AI: Moving From Hype to High-Impact

Bill Benton, co-founder of GAINSystems, explains how companies can escape the hype surrounding artificial intelligence, and embrace the technology in a way that brings real value to supply chains.

The proper use of AI starts with defining the specific problem to be addressed, using an “impact matrix,” Benton says. The focus should be on identifying events, and possible reactions to them, that result in the biggest effect on the enterprise.

The goal should be to emphasize speed of deployment and a realistic approach to what’s possible. The action in question is usually “unglamorous,” Benton says, involving relatively modest goals. Objectives such as fully autonomous planning or near-perfect demand prediction should be seen as “later-stage.”

In theory, he says, one could review thousands of possible scenarios and data points, but the wiser plan of action is to narrow those down with the help of AI in order to initiate discrete and relevant actions that complement, not replace, human activities within the organization.

The biggest misconception around the use of AI today is the failure to tell the difference between “science projects and deployable solutions,” Benton says. Some 83% of highly ambitious deployments either fail outright or achieve less than half of the goals set from the start.

The mistake, Benton says, “is not discerning the probability of success in an authentic way, but going for the maximum high.” Instead, supply chain executives should seek “quick wins,” then “move incrementally so that the last success funds the next.”

AI can have an “immediate and profound” impact in creating distinct scenarios that are most relevant to the organization. To a certain extent, its conclusions are based on prior experience, but the technology can also assign probabilities to events that are unprecedented, and for which the business has no planned response.