Ninety-five percent of enterprise agentic AI pilots never deliver ROI, according to MIT’s 2025 research. The gap isn’t ambition — it’s the jump from a model that performs well in a sandbox to an agent trusted to act on the plant floor, inside the systems your team already runs on.In this session, Supply Chain Now hosts Scott Luton and Kim Reuter sit down with DataRobot’s Vika Smilansky and Marc Amarillas to break down what that jump actually takes, and what it looks like solved, with agents orchestrated across the value chain instead of isolated in silos:

• Supply chain: multi-tier supplier risk and tariff exposure caught before a single-source failure becomes a line-down event
• Plant operations: bottlenecks and downtime caught before they hit the shift report
• Aftermarket: warranty leakage and reactive dispatch turned back into retained margin

The conversation is grounded in a real use case: a tariff signal that put $2.5 billion in revenue at risk, resolved in 72 hours instead of weeks, with a human signing off at every step.

What you’ll learn:

• How agent workforces catch multi-tier supplier risk and tariff exposure before a single-source failure turns into a line-down event
• How plant operations agents catch bottlenecks and downtime before they hit the shift report
• How aftermarket agents turn warranty leakage and reactive dispatch back into retained margin
• What separates a production-grade agent workforce from a pilot or a dashboard

Hosted by Scott Luton and Kim Reuter of Supply Chain Now.