Executive Summary
Artificial intelligence is becoming an established part of procurement, but organizations are taking a measured approach to how broadly they deploy it and how much autonomy they give it. This ProcureCon Insights report examines how procurement leaders are using AI, where it is delivering measurable value and what stands in the way of broader adoption.
Among 100 procurement leaders surveyed in North America, 75% say AI is partially integrated across several procurement functions, while 14% report it is highly integrated across most or all workflows. The most commonly cited areas of measurable value are process efficiency and cycle-time reduction at 31%, and improved data visibility and reporting at 30%. Cost savings and spend reduction rank much lower at 5%.
The findings also identify barriers shaping the next phase of AI in procurement. Forty-two percent of respondents cite data privacy, security and compliance concerns among their top barriers to implementing or scaling AI, while 35% cite difficulty integrating AI with existing procurement platforms. Resistance to change and a lack of AI skills or expertise each register at 31%.
As organizations explore agentic AI and greater procurement automation, most respondents continue to favor human oversight. Sixty-six percent say AI should handle classification and routing while people handle everything else.
The report examines what these findings mean for AI investment, governance, skills development, and the balance between automation and human expertise.
Read the paper now.
Frequently asked questions
Among respondents, 75% report AI is partially integrated across several procurement functions, 14% say it is highly integrated, and 11% are piloting it in one or two areas.
Respondents most often cite process efficiency and cycle-time reduction, at 31%, followed by improved data visibility and reporting, at 30%. Cost savings and spend reduction rank at 5%.
Forty-two percent cite data privacy, security and compliance concerns among their top barriers, followed by integration difficulties at 35%. Resistance to change and lack of AI skills each register at 31%.
Sixty-six percent say AI should handle classification and routing while humans handle everything else. Another 28% support end-to-end management of routine requests, with exceptions escalated to people.
Better compliance and audit trails rank first, selected by 44% of respondents, followed by measurable cost savings at 34% and reduced supplier and supply chain risk at 17%.