Only 5% of procurement organizations have reached wide-scale implementation of GenAI, a sharp contrast to the near-universal interest recorded in the 2026 CPO Annual Pulse Report – State of GenAI in Procurement from EFESO Management Consultants. While enthusiasm is high, EFESO notes that the field is moving into a more selective phase of investment, one shaped by data governance, operating-model constraints, and clearer ROI expectations.
Experimentation Accelerates, but Scaling Hits Structural Limits
The report shows a broad wave of experimentation: 75% of organizations are still in early exploration or pilot stages, while just 20% have partially deployed GenAI into core workflows. This execution gap mirrors broader enterprise trends, recent data shows that AI budgets are expanding, but deployment rigor is tightening, especially as companies strengthen controls around data risk and model accuracy.
Adoption is sharply uneven across company size. In large enterprises, 83% of respondents report using GenAI regularly at work, supported by dedicated AI teams and structured digital roadmaps. SMEs, by contrast, show just 38% regular use, with efforts often spearheaded by individuals rather than integrated into enterprise-wide systems. This aligns with broader technology investment patterns, where smaller companies face higher barriers related to tooling, governance, and foundational data quality.
GenAI’s strongest foothold sits in source-to-contract tasks where the data environment is dense and structured. Contract analysis and summarization lead adoption at 69%, followed by sourcing intelligence (61%) and RFx automation (55%). These areas offer rapid productivity gains with relatively lower integration risk, a pattern also reflected in other functions adopting GenAI for document-heavy workloads.
More advanced applications such as AI-supported negotiation remain constrained, with only 35% seeing meaningful value today. EFESO attributes this to the deeper data integration and governance required for models to support or simulate real negotiation dynamics.
Value Delivery Remains Uneven as Trust and Skills Slow Deployment
Despite rising usage, only 34% of organizations say GenAI value matches expectations, while 46% are partially satisfied and 20% express disappointment. Concerns about data reliability (68%) and confidentiality and regulatory compliance (67%) continue to weigh heavily on scaling decisions, especially as companies navigate emerging AI governance requirements in multiple markets.
Skills shortages remain another structural barrier. More than half of respondents cite capability gaps or inconsistent data readiness as obstacles to moving beyond pilots. EFESO interprets this not as backsliding but as evidence of maturing procurement functions that are becoming more selective about where GenAI belongs.
Broader market signals reinforce this pivot: enterprise AI spending is still rising, but CFOs across industries are tightening their scrutiny of payback periods, integration costs, and model reliability. Procurement teams appear to be responding in kind, shifting from broad experimentation toward targeted deployment where value can be clearly demonstrated.
Why Selective Scaling May Redefine Procurement Architecture
One overlooked dynamic is how selective GenAI deployment is beginning to reshape the architecture of procurement platforms themselves. As vendors incorporate model governance, audit trails, and AI-native workflows into their roadmaps, the technology is influencing how systems interconnect rather than simply augmenting existing processes. Recent product announcements across major suites show a shift toward embedded model oversight and traceable decision flows, features that were peripheral even a year ago. That trajectory suggests procurement teams may soon evaluate platforms not only on functional depth, but on how effectively they operationalize transparency and control around AI-driven decisions.