Procurement teams are on the front lines of today’s global supply chains. They’re expected to react to fast-changing economic and geopolitical shifts that impact sourcing strategies, even as they struggle to make sense of mountains of data relating to regulation, purchasing patterns and intensifying customer-service demands.

Given the immensity of that challenge, it’s no surprise that artificial intelligence is playing an increasingly important role in global supply chain management, especially for the procurement function. But is AI having the desired effect on that discipline? And, is it poised to supplement the vital role of procurement professionals, or replace them?

Two analysts from Gartner addressed those questions at the consultancy’s recent Procurement Conference in San Diego.

Senior director analyst Fareen Mehrzai noted a 360% increase in procurement roles requiring AI skills over the last three years. Yet along with that surge comes higher levels of anxiety among executives about what the technology will actually do to their jobs.

Are those fears justified? “It’s the adult version of the bogey monster,” said Mehrzai. “Many are pretending it’s not there, hoping it’s not coming, or just going through training exercises.”

She said companies need to create an environment of “psychological safety” — one in which managers feel comfortable sharing opinions, soliciting feedback and even making mistakes in their assessment of AI’s advantages and potential drawbacks.

Mehrzai recommended a three-pronged strategy for coping with the coming of AI to procurement:

Deploy AI human-first. “You are in control of the 'how,'” she said. The goal should be to enable to people to work more efficiently. Staff needs to adopt an “abundance mindset” that focuses on the possibilities of AI solving complex problems while boosting the creativity of human teams. “The tone you take determines what you’re going to get from AI,” Mehrzai said.

Diagnose the true root cause of AI fear. Questions about the technology’s impact must be addressed in “an ethical and thoughtful way,” accounting for the “deep-seated emotions” of managers who understandably wonder whether their skills are still relevant in a world of AI. Mehrzai suggested setting up AI “learning zones” that “reduce the cognitive strain associated with trying new AI tools.” In addition, she said, companies should talk about what’s not changing in human roles and responsibilities.

Co-create “high-touch” strategies. Mehrzai cited the examples of computer maker Lenovo, which engaged managers in “active listening” exercises and emotional-intelligence coaching as strategies for easing the adoption of AI; and the Australian law firm MinterEllison, which defined five maturity levels of AI adoption among staff: those who avoid, understand, explore, adopt and champion the technology. By customizing learning in line with each individual’s propensity for acceptance, the firm increased the number of employees using AI weekly by 50%.

Ryan Polk, Gartner’s chief of research for procurement, laid out four questions that leaders must ask about the potential for AI to “reinvent” procurement:

What’s holding us back from achieving this vision? The number-one reason for failure of an AI initiative, Polk said, is poor data management. To work effectively, AI agents require both “semi-structured” data (metadata, market research, supplier performance and contract details) and unstructured data (document repositories, images, videos, sound and other intelligence obtained from social media).

When AI becomes ubiquitous, where does competitive advantage come from? Polk’s answer: “The human being.” Staff should be focusing on edge cases and exceptions, while AI takes on more standardized work, “where the data is known or readily available.” New staff competencies will emerge, he said, as AI agents shift the skills that are required in procurement roles. Humans will be expected to acquire expertise in such areas as data literacy, prompt engineering, inductive analysis, and customer-journey mapping. As AI takes hold, they’ll have less of a role in deductive analysis, process execution, negotiation, and financial and legal acumen.

How do we influence buying decisions in the AI-augmented era? Increasingly, Polk said, corporate buying journeys are starting with AI, with 80% of enterprise research teams expected to use the technology by 2028.  “You have to actively govern the data [and] plug it into all internal stakeholders,” he said.

What second-order effects will AI investments have on the economy that I need to be aware of? “Every digital decision has a physical footprint somewhere in the economy,” Polk said, citing the growing requirements for power, transformers, storage and other physical demands of an AI buildout at scale. “We will have to think in future about capacity constraints, limited alternatives, regional ecosystems, supply assurance and collaboration,” he said. “We can’t negotiate a path that doesn’t exist.”

Polk recommended four actions for procurement teams to take, as they look to adopt AI: Identify strategic constraints within the organization; invest in the technology early; build deeper partnerships, including by measuring and managing supplier development; and — echoing the theme that rings out with every discussion of AI today — “define the role of the human.”

Next: AI in Manufacturing