Procurement should start with high-volume, repeatable work that follows clear rules and uses reasonably reliable data. Prioritize tasks such as purchase-to-pay processing, invoice handling, classification, intake routing, spend analysis, and contract-data extraction, then keep people accountable for negotiations, supplier relationships, complex sourcing decisions, and material exceptions.

At ProcureCon Indirect West, Stephen Osmun, Director of Procurement at GoFundMe, argued that the first filter should be the nature of the work: "Is this work rules-driven or is this judgment-driven?" He identified structured, predictable purchase-to-pay activity as a strong starting point because it can reduce manual effort and release capacity for supplier relationships, category strategy, and other higher-value work.

The 2026 ProcureCon Insights report An Achievable Future for AI in Procurement found that 31% of senior procurement leaders identified process efficiency and cycle-time reduction as AI's most measurable source of value, while 30% identified improved data visibility and reporting. The results suggest that early AI programs should be evaluated first on speed, insight, and usability rather than promises of immediate savings.

For CPOs, procurement transformation leaders, and centers of excellence, the practical next step is to create a short list of candidate tasks and score each one for volume, repeatability, data quality, business risk, and required human judgment. Begin with one bounded use case, establish a baseline for time, quality, exceptions, and adoption, then decide whether to expand based on evidence.

Join us at ProcureCon Indirect West to compare practical approaches for applying AI to routine procurement work while keeping human judgment at the center of consequential decisions.