AI in Procurement: Returning to the Core of Negotiation
Generative AI is rapidly moving from a future promise to a present-day reality in procurement. Yet despite the momentum, most organizations are still in an early experimentation phase rather than full-scale adoption.
In a recent article published in Germany in collaboration with Beschaffung aktuell, EFESO experts Dr. Kai Dresch and Julian Tummann explore how AI is reshaping procurement negotiations and where its real value lies today.
Based on insights from EFESO’s 2026 GenAI Procurement Pulse Report, which draws on interviews with 50 Chief Procurement Officers across Europe, the findings show a clear gap between hype and operational reality: while AI dominates the agenda, only a small minority of organizations have scaled its use, with the vast majority still running pilots.
From Hype to Operational Reality
As Kai Dresch highlights, AI is already proving its value, but primarily in specific, data-intensive areas rather than as a fully integrated end-to-end solution. The biggest barriers remain structural: fragmented data landscapes, legacy processes, and change management challenges that slow down adoption.
Julian Tummann emphasizes that success depends less on the technology itself and more on how it is embedded into procurement ways of working, especially ensuring stakeholder alignment and trust in AI-supported decisions.
Where AI Delivers Real Value Today
The article identifies clear use cases where AI is already creating measurable impact across the procure-to-pay process, particularly in:
- Contract analysis and summarization
- Sourcing intelligence and market analysis
- RFx automation and supplier insights
In sourcing, for example, AI can help process market, supplier, and ESG data in near real time, significantly increasing transparency and decision quality.
Negotiations: AI as a Co-Pilot, Not a Replacement
A central focus of the article is the role of AI in negotiations. Rather than replacing human judgment, AI strengthens it across three key phases:
- Preparation: analyzing historical pricing, supplier performance, and market trends to support strategy
- Execution: providing real-time data and structured decision support during negotiations
- Post-negotiation: improving documentation, tracking commitments, and enriching future analysis
As the authors underline, AI should be seen as a “co-pilot” in complex negotiations enhancing clarity and speed, but not substituting human experience and judgment.
A Differentiated Approach to Negotiation Types
The article also proposes a practical framework distinguishing negotiation types by complexity and strategic value:
- Speed: highly standardized negotiations that can be largely automated
- Efficiency: negotiations where AI supports analysis, benchmarking, and scenario modeling
- Performance: strategic negotiations where AI informs decisions but human expertise remains critical
- Break-up: low-value, high-complexity areas that should be deprioritized or restructured
Conclusion: intelligent orchestration over full automation
The key takeaway is clear: the future of procurement is not fully autonomous negotiation, but intelligent orchestration.
AI will increasingly handle data-heavy tasks, accelerate analysis, and improve transparency. However, in complex and strategic negotiations, human judgment remains irreplaceable.
As Dresch and Tummann conclude, the real differentiator will not be who adopts AI fastest, but who succeeds in combining technology, data, processes, and people into a coherent “negotiation orchestra.”
Lesen Sie den vollständigen Artikel hier: Beschaffung aktuell
The 2026 CPO Annual Pulse Report
EFESO’s 2026 CPO Annual Pulse Report reveals a striking gap between experimentation and execution: while nearly all procurement organizations are piloting GenAI, only 5% have successfully scaled it across the function.
Drawing on insights from 50 European Chief Procurement Officers, the report explores why enthusiasm has not yet translated into widespread adoption.
As procurement leaders move from hype to hard choices, the study uncovers what separates isolated pilots from disciplined, value-driven deployment and why scaling GenAI is proving far more complex than expected.