Individual gains that do not extend to the organization
At the Gartner Supply Chain Symposium/Xpo, held in Barcelona in May 2026, a survey of 101 CPOs revealed a figure worth examining closely: only 36% of procurement executives say they are truly confident in their ability to rethink their organization’s roles and processes around generative artificial intelligence.
This figure does not call into question the adoption of AI. It is now widely used in most procurement departments: accelerated writing, document summarization, and initial drafts of RFPs. It raises another, more structural question: why do very real individual gains not translate into collective performance for the procurement function as a whole?
An individual gain that doesn’t trickle down to the organization
Gartner calls this phenomenon the AI productivity paradox. The observation is simple to state but harder to address: when AI automates part of a task, the user saves time and improves quality. However, the organization itself does not automatically reap this benefit. Without a review of roles, workflows, and performance metrics, the gain remains limited to the individual level.
This is a point that is often underestimated in discussions about AI applied to procurement: technology only generates collective value if the organization is designed to capture it. A faster tool alone does not make a procurement department more effective.
A Method, Not a Promise
Gartner’s analysis goes beyond mere observation. It proposes a concrete approach that any procurement department can implement without waiting for a major transformation project:
- Mapping actual tasks. Before discussing AI, we need to compile a precise inventory of the activities carried out by procurement teams as they are performed today.
- Distinguish between what remains a human task and what can be natively automated. Negotiation, strategic decision-making, and supplier relations are skills that are difficult to replace. Data collection, information structuring, and initial analysis, on the other hand, can be largely handled by AI.
- Revise performance metrics. Traditional metrics, which focus on volume produced per unit of time, are becoming unsuitable for an AI-powered procurement function. Gartner recommends replacing them with value-based metrics: contribution to financial results, complexity actually absorbed by the organization, and new types of production made possible.
What this means in practice for a procurement department
This assessment shifts the focus of the debate. The question is no longer whether to deploy AI in procurement; that decision has largely been settled. The question has become an organizational one: who in the procurement department will take charge of redesigning roles, workflows, and performance metrics?
This shift is part of a broader movement to reposition the procurement function. Long driven by cost and operational efficiency metrics, the function is increasingly expected to demonstrate its ability to create measurable value, manage complexity, and contribute directly to the company’s bottom line. The productivity paradox identified by Gartner is a clear illustration of this: it is not just a matter of equipping buyers, but of redefining what is expected of them.
For a procurement department, the immediate priority is therefore not to increase the number of AI use cases, but to carry out this preliminary work of mapping and reviewing roles. It is this step,more than the tool itself,that determines the organization’s ability to transform individual gains into collective performance.
This is precisely what has guided our approach at Oalia from the very beginning: a platform has value only if it is designed to structure this collective effort, not merely to equip isolated users. It is this distinction, more than the technology itself, that separates a tool we use from an organization that makes progress.
Source: Gartner, presentation at the Supply Chain Symposium/Xpo, Barcelona, May 2026, reported by Supply Chain Digital.