Agentic AI Transforms Procurement: A Call for Enhanced Risk Management

The Dawn of Agentic AI in Procurement

Procurement, traditionally a significant source of organizational risk, is undergoing a profound transformation with the advent of agentic AI. Unlike previous analytical AI, which merely presented data, agentic AI actively emulates human judgment, executes multi-step tasks, and continuously learns to improve outcomes. This new wave of AI acts as a 'digital colleague,' capable of analyzing supplier bids, tracking market indices, flagging cost deviations, and even preparing negotiation strategies autonomously.

Real-World Impact and Emerging Risks

McKinsey highlights compelling examples of agentic AI's impact:

• A tech company used linked AI agents to rebuild its external services sourcing strategy, integrating spend and market data to generate real-time insights. This led to significant savings: 12-20% in contact center operations and 20-29% in business process outsourcing (BPO) and financial services spend.
• A chemicals company is piloting AI agents for autonomous sourcing in consumables, automating tender preparation, supplier prequalification, and bid analysis. This has boosted procurement staff efficiency by 20-30% and increased value capture by 1-3%.

However, these advancements introduce critical new risks that demand attention:

• Data Reliability: How can organizations ensure the AI agents are using complete, reliable, and up-to-date data for their analyses and recommendations?
• AI Learning and Adaptability: If agents are 'learning' and modifying themselves, how can their continued reliability and adherence to organizational policies be guaranteed?
• Decision-Making Transparency: Can we rely on agents to consider all options and related risks, especially when evaluating complex factors like vendor performance?
• Inter-Agent Communication: What are the implications when an organization's AI agents interact with those of suppliers or other external entities? Whose interests are prioritized in such exchanges?
• Development and Maintenance: Who develops and maintains these agents, and how are changes tested and validated, especially if other agents are involved in their maintenance?

The Imperative for AI Governance

As AI agents increasingly reshape business processes, operational risks will inevitably change. This necessitates a proactive approach to AI governance. Organizations must ask:

• Are we prepared for the evolving risk landscape introduced by agentic AI?
• Is effective and efficient AI governance in place, or is AI being implemented in fragmented silos without proper oversight?
• Are lessons learned from AI implementations being shared and integrated across the organization to inform future strategies and risk mitigation efforts?

The shift to agentic AI is not just a technological upgrade; it's a fundamental change in how businesses operate, demanding a corresponding evolution in risk management and internal audit practices.