Before scaling AI, procurement must improve the accuracy, consistency, ownership, and accessibility of its supplier, spend, contract, and transaction data. Teams should also connect relevant data across ERP, sourcing, contract, risk, and intake systems, then establish clear rules for who maintains it and how it can be used.

At ProcureCon Indirect West, David Rodriguez, Director of Strategic Sourcing at Tarsus Pharmaceuticals, described how AI can combine ERP data, market information, supplier intelligence, and sourcing inputs to identify opportunities and accelerate early work. But he also cautioned that, "Your data with AI is never going to be clean either," underscoring that technology can improve analysis without removing the need for human review and disciplined data management.

The 2026 ProcureCon Insights State of Procurement Report found that 48% of leaders see data quality or infrastructure issues as a top internal obstacle to gaining measurable value from procurement AI. This result suggests that many teams need to address data foundations before expecting broad, trusted results from new tools.

Procurement analytics leaders, CPOs, and CIOs should not begin by attempting to clean every enterprise dataset at once. Instead, they should identify the data needed for their first AI use case; assign owners for supplier masters, spend categories, contracts, and performance data; standardize key definitions; correct the highest-impact errors; and document where human review remains necessary when information is incomplete or conflicting.

Join us at ProcureCon West to explore how procurement leaders are improving data quality, integration, and governance before expanding AI across the function.