By its very nature, high-tech should be more advanced than other business sectors in the adoption of artificial intelligence for supply-chain planning. But that’s not necessarily the case.
“It’s one of those things that everybody assumes everybody is doing,” says Brian Tessier, vice president of global supply-chain innovation with Schneider Electric. “But nobody really knows what it means. It’s five miles wide and five miles deep.”
The general awareness and application of A.I. is on the rise, notes Tessier. Smart devices such as intelligence speakers are becoming increasingly common in homes. In high-tech, there are “pockets of best practice, but it’s very variable right now.”
There are two distinct approaches to A.I. today, says Madhav Durbha, vice president of industry strategy with Kinaxis. One is the traditional “purist’s” ambition of designing computer programs to mimic the human brain. Another is a more practical effort that involves an application-specific use of A.I. The latter approach doesn’t necessarily seek to replace human planners with machines. Instead, it works to apply the technology to defined use cases within an organization.
Joe Lewis, senior manager of Deloitte Consulting, agrees that many companies are pursuing what has been referred to as “narrow A.I.” The near-term goal, he says, is to combine A.I. and machine learning with human experts. Nevertheless, there is a debate as to whether demand planners will ever reach the point where people are taken out of the loop altogether, and computers are able to outperform the human brain.
For now, “we are close to achieving specific, scaled-down tasks,” Lewis says. A.I. is increasingly being applied to supply-chain management to identify trends, monitor supplier quality and perform inventory optimization.
The problem for A.I., says Durbha, is that supply-chain planners typically must operate with imperfect and inadequate data. “That’s where machine intelligence stops, and human creativity needs to come into the picture.”
Tessier says A.I. and machine learning are well equipped to assume the “drudgery” of basic planning processes. The challenge for a company like Schneider Electric is that it must align processes and definitions from multiple organizations in the event of a merger or acquisition. In such instances, it’s could be valuable to apply aspects of “non-traditional” A.I. – even those that aren’t considered applicable to business. “Learning through observation is where the majority of these systems are today,” he says.
It’s somewhat ironic that the effort to replace people with machines invariably leads to the question of human talent – specifically, where it will come from in the future. Qualified individuals will need to demonstrate both “strong analytical skills and business acumen,” Durbha says. College graduates with double majors in mathematics and business are ideal candidates for the supply-chain planning job of the future. The question is whether there will be enough of those individuals to go around.
Lewis believes consulting organizations such as Deloitte will come to rely on multiple individuals with specialized skills, working together as a team. “I’m optimistic about the future of talent,” he says. “We’re starting to see more broad-based people coming down the pike.”
