New Research by Ardent Partners and Ivalua Reveals Gap between AI Ambition and Execution in Procurement

Research finds 67% of procurement teams have high AI ambitions but are not ready to scale, with 59% citing poor data quality/structure/availability as the top blocker and 34% lacking formal AI governance. Only 23% report successfully scaling “AI-First” procurement via strong data foundations and embedded governance, while 47% target cognitive augmentation in a human-agent model. The data suggests near-term deployment gaps (e.g., contract lifecycle management planned 56% vs 14% in use), tempering expectations for near-term, broadly scalable ROI.

Analysis

This is less a demand-surge story than a budget reallocation toward data plumbing. In procurement, AI monetization will accrue first to platforms already embedded in workflows and master data, while bolt-on copilots face the classic enterprise trap: pilots are easy, production is hard, and the approval layer slows conversion.

The second-order winners are ERP/procurement suites, integration layers, and consulting/service firms that clean up data and define governance. The losers are point solutions and pure AI narrative names that need a seamless data model to prove ROI; in a control-heavy function, black-box concerns should lengthen sales cycles and keep contract values small until vendors can show auditability and human-in-the-loop controls.

Consensus is likely overestimating how quickly procurement AI becomes revenue. The near-term catalyst is commentary, not earnings: expect vendors to talk up pilots over the next 1-2 quarters, but the real test is 2H26 booking conversion and module expansion. Falsifiers include sustained acceleration in procurement-specific ARR or evidence that buyers will accept autonomous workflows without manual approval, which would compress the governance premium.

Key Decisions for Investors

• Long ORCL / short AI as a 3-6 month pair: Oracle should benefit from the unified-data/governance requirement, while standalone AI names remain exposed to slow enterprise conversion. Target 2:1 reward/risk if procurement AI spend stays pilot-heavy.
• Long SAP vs short a basket of pure-play AI software names (AI, SOUN) into the next 1-2 earnings cycles. Thesis: procurement is likely to buy architecture, not models. Cut the short if any of the AI names show durable enterprise ARR reacceleration.
• Use ACN and IBM as secondary longs on weakness: they are positioned to capture the services work required to normalize data, build controls, and operationalize AI in procurement. Best entry is on post-earnings dips or after management guides to higher transformation demand.
• No direct trade in GAP/WWRL unless subsequent filings show explicit procurement-software exposure; treat these as watchlist names only. The article is an execution-readiness survey, not a hard revenue catalyst.