EFESO’s 2026 CPO Annual Pulse Report reveals why GenAI pilots are in every single place, however disciplined scale stays uncommon
Generative AI is not a novelty in procurement. It’s broadly examined, actively debated, and nonetheless not often scaled. In response to the 2026 CPO Annual Pulse Report – State of GenAI in Procurement, revealed by EFESO Administration Consultants, solely 5% of procurement organizations have achieved widespread deployment of GenAI, regardless of near-universal experimentation. Primarily based on discussions with 50 European Chief Procurement Officers, the examine reveals a transparent shift from hype to disciplined, selective execution.
Pilots in every single place. Scale virtually nowhere.
The Pulse confirms that GenAI is firmly on procurement’s agenda, however large-scale deployment continues to be restricted. 75% of procurement organizations stay in experimentation, with 40% nonetheless in early exploration and 35% working pilots. Solely 20% report partial deployment, revealing a rising hole between ambition and execution.
“What we’re seeing isn’t a scarcity of curiosity, it’s an increase in self-discipline,” says Gael Sandrin, Principal at EFESO France, main the CPO annual pulse. “Procurement leaders are not asking whether or not GenAI works. They’re asking the place it really works, at what price, and underneath which situations.”
Additionally Learn: AiThority Interview With Claire Southey, Chief AI Officer at Rokt
The Pulse additionally highlights a widening digital divide. 83% of respondents in giant enterprises report common use of GenAI at work, in contrast with simply 38% in SMEs, reflecting variations in budgets, devoted groups, and structured AI applications, whereas smaller organizations proceed to depend on fragmented, individual-led experimentation with restricted integration into core processes.
Early GenAI worth is clearly seen, however concentrated. Contract evaluation and summarization lead adoption (69%), adopted by sourcing and market intelligence (61%) and RFx automation (55%). These productivity-driven, source-to-contract use circumstances profit from excessive knowledge density and decrease integration danger. Extra complicated functions lag behind: solely 35% see worth at this time in AI-supported negotiation, which requires deeper knowledge integration, governance, and extra superior system design.
Expectations reset as belief and execution take middle stage
Regardless of widespread testing, satisfaction stays combined. Solely 34% of organizations report being happy with GenAI worth relative to preliminary framing and funding, whereas 46% report partial satisfaction and 20% specific disappointment. Belief-related points weigh closely on scaling selections, with knowledge reliability (68%) and regulatory compliance and confidentiality (67%) rating among the many high issues. Expertise shortages (57%) and knowledge high quality limitations (55%) additional constrain progress.
“2026 is not going to be the 12 months of GenAI in every single place,” provides Kenneth Sievers, Companion and world head of Procurement at EFESO. “It will likely be the 12 months of strategic readability. The aggressive hole will widen between organizations that industrialize selectively, and those who stay caught in perpetual pilots.”
Somewhat than signaling a slowdown, EFESO interprets these findings as an indication of rising maturity. Procurement organizations are transferring away from broad experimentation towards selective industrialization, specializing in clearly outlined use circumstances the place feasibility, knowledge readiness, financial credibility, and governance align.
The 2026 CPO Annual Pulse Report – State of GenAI in Procurement, EFESO’s second annual examine, offers an in-depth view based mostly on qualitative discussions with senior procurement leaders throughout Europe, between November and December 2025, on how they’re adopting generative AI, the place worth is rising, and why disciplined execution is now the defining problem.
Additionally Learn: The Physics of Intelligence: Can AI Techniques Develop an Inner Mannequin of Actuality?
[To share your insights with us, please write to psen@itechseries.com]
