Share with your COO
Zycus is betting that procurement’s AI moment is a governance crisis, not a technology gap. At its Agentic AI Procurement Summit 2026, the company drew 1,500+ procurement leaders across 50 countries and released two research reports alongside live demos of its Merlin platform. The headline numbers are striking: 58% of procurement leaders expect agentic AI to materially impact their organizations within twelve months, while fewer than half of CPOs feel confident they can monitor and control it once deployed.
What this means for your business
If your procurement function is still running pilots, the twelve-month window in that Hackett Group data is the number to sit with. The gap between pilot and enterprise-wide deployment isn’t a technology problem at most organizations, it’s a control problem. Fewer than half of CPOs feeling confident they can monitor agentic AI, meaning AI systems that act autonomously across multi-step tasks without human approval at each step, tells you the bottleneck isn’t capability. It’s the absence of a governance architecture that operations leaders actually own.
Forrester’s report title, “Don’t Delegate AI,” is doing real work here. The argument that the CPO, not IT and not the vendor, should own AI strategy direction and guardrails is correct, and Zycus, which sells the platform those guardrails would run on, has an obvious interest in CPOs believing they need to get more deeply involved rather than defer. That incentive tilts the framing toward urgency. But the underlying claim holds regardless: when autonomous agents are negotiating supplier contracts on limited human input, the person accountable for procurement outcomes has to own the decision rules, not just the results. Delegation to IT produces systems optimized for deployment; delegation to vendors produces systems optimized for renewal.
The leading indicator to watch isn’t adoption rate, it’s where AI strategy ownership actually lands in the next org-design cycle. If your COO or CPO hasn’t been assigned formal accountability for agentic AI guardrails by the time your next procurement platform contract comes up for renewal, the vendor will fill that vacuum by default. That’s the budget conversation worth having now, not after the pilot scales.
Concept deep-dive: Agentic AI
Agentic AI refers to AI systems that don’t just answer questions but take sequences of actions autonomously to complete a goal, think of it as the difference between a calculator and an employee who handles the whole task from intake to completion without checking in at every step. In procurement, that means an agent could run a supplier negotiation start to finish with minimal human sign-off. The governance question is exactly who defines the boundaries of that autonomy before it runs.
Based on reporting from Zycus Concludes Agentic AI Procurement Summit 2026; Forrester and Hackett Group Research Released | Press Releases, originally published 2026-06-09 03:00:00.
