Your lengthy career in procurement has consisted largely of being beaten up by top management over the need to constantly cut supplier costs. Along comes digital technology to make your job easier — or take it away.
That, at least, is the fear of procurement professionals in the age of artificial intelligence and machine learning. Considering all of the factors that go into supplier selection and maintenance, doesn’t it make sense to hand over the task to a system that can crunch more data than a hundred humans?
And the volume of data continues to grow. Procurement managers today must draw on multiple sources of intelligence, from suppliers themselves as well as independent financial data, news reports, third-party services and social media. Says Sammeli Sammalkorpi, co-founder of procurement analytics software vendor Sievo: “Procurement organizations haven’t yet learned how to tap into these different kinds of data.”
AI and machine learning would seem especially well-suited to the challenge. But fears that they’ll take over the job entirely seem unfounded. When it comes to effective procurement and supplier management, the foreseeable future is one of collaboration between human and machine.
“I really believe that the role of AI is not to replace humans,” says Sammalkorpi. “Machine learning and AI can propose findings, but you still need to adjust to what’s relevant, and what’s not.”
Machines are highly effective for narrow applications, involving well-defined problems. What’s more, they’re always “on” and entail a minimal cost of operation, when compared to the salary and benefits of a human worker.
Comes time to take action, however, and the flesh-and-blood manager needs to step in. Simply put, humans are still better at making the final judgment about key suppliers. For the time being, at least, “machine learning is still not reliable enough to drive decisions,” Sammalkorpi says. (There’s another reason for keeping humans in in the loop, he adds: they need to retain responsibility for making choices so that they don’t end up blaming the machine for wrong ones.)
In the early stages of machine learning, the technology can likely take over certain elements of the supplier contract. As it improves, however, buyers will find themselves relying on the system for an increasing number of tasks, if not the final decision on supplier vetting and selection.
Because AI and machine learning depend heavily on complex algorithms, companies might assume that they need to hire a expensive team of data scientists to run the system and make sense of its conclusions. In Sammalkorpi’s opinion, that’s not the case for the procurement department. He believes that kind of expertise is better obtained from an outsourced provider.
“Even if they thought it was a good business case,” he says, “we’re not seeing a lot of organizations capable of retaining that talent in house. You still need data scientists, but I don’t think the procurement organization is the right place for them.”
Predicting future outcomes is likely to be as tough for a machine as for a human — in other words, impossible. But machine learning is good at quickly revising forecasts and action plans in real time, to reflect actual purchasing patterns, Sammalkorpi believes.
All of this assumes, of course, that procurement can smoothly incorporate the new technology into its operations. But new research from Forrester Consulting, commissioned by procurement platform vendor Ivalua, suggests that this is far from the case. Employing a “digital maturity index” to assess companies’ progress in adopting procurement technology, Forrester found that most are “significantly overestimating” their level of maturity in that respect.
Sixty-five percent of companies surveyed considered themselves to be “advanced,” but only 16 percent had the requisite level of digital maturity in their procurement organizations to justify that assessment.
One problem is that many companies make poor initial choices in selecting procurement technology. In the Forrester study, 82 percent had switched or were considering switching tech providers, citing as reasons poor levels of supplier onboarding and user adoption.
Moreover, the adoption period was excessively long. Just 17 percent of organizations were able to onboard new suppliers in less than a month, and 59 percent were taking between one and three months for each supplier.
“Procurement leaders have the opportunity to deliver a true competitive advantage for their organizations,” David Khuat-Duy, corporate chief executive officer of Ivalua, said in a statement. “Digital transformation is critical to success, but requires a realistic assessment of current maturity, a clear vision for each stage of the journey and the right technology.”
All of which suggests that technology in the form of AI and machine learning is a long way from jettisoning humans from the procurement function, even as it promises to enhance operations when properly assessed and implemented.