Human accountability remains central to self-driving S&OP

At Reuters Supply Chain USA 2026, Sheri Hinish moderated a panel examining whether S&OP can evolve into a self-driving supply chain. The discussion suggested that technology is not the primary obstacle: people, process, trust and accountability remain the harder constraints.

Diego Faria described a confidence-scoring approach in which planners validate AI-generated safety-stock recommendations over several cycles. Autonomy expands only after the recommendation repeatedly earns human confidence, while the planner remains accountable for approving the decision.

Chinmay Bhatt argued that technology represents only part of implementation success. His sequence—process, people, platform and then AI—places change management and operating-model discipline ahead of automation.

Prashant Gopalan compared autonomous planning with the gradual adoption of self-driving vehicles. In Carpenter Technology’s compliance-sensitive aerospace environment, AI agents are being developed around institutional knowledge while certification and human oversight remain mandatory.

Mark Sawaya identified end-to-end network optimization as the larger value opportunity. He also questioned whether traditional monthly S&OP cycles remain appropriate when AI can compress the sequence of collecting data, analyzing options, meeting and deciding.

Operational perspective: autonomous planning should be earned decision by decision. Each use case needs a defined owner, confidence threshold, validation period, escalation rule and failure response. A system may generate safety-stock or network recommendations, but accountability for service failures, compliance breaches and operational trade-offs cannot remain ambiguous.

Organizations should begin with reversible, bounded decisions and retain the approved recommendation, evidence, approver and resulting outcome. More consequential actions require reconciled execution through APS and ERP systems, audit trails, manual overrides and explicit governance over when autonomy may be expanded or withdrawn.