In this talk, Sylvain Chatonnier of SAP presents the vendor's autonomous supply chain strategy, structured around five pillars: planning, product definition, manufacturing, logistics, and industrial asset management — a full cycle SAP calls "design to operate."
At the heart of this strategy is Joule, SAP's AI copilot, which provides assistants for each major domain, themselves backed by a multitude of operational agents. The speaker details the technical foundations needed to make this system reliable: the SAP Knowledge Graph, which structures relationships between business objects (customers, orders, suppliers), and an agent governance tool compared to an "HR system for agents," used to define each agent's rights and privileges.
Several demos illustrate increasing levels of agent sophistication: automatically resolving a stock shortage in long-term planning, launching a manufacturing order via a simple prompt, a multi-agent analysis following a bill-of-materials change, and a predictive maintenance scenario triggered by IoT sensors on an industrial pump. The session closes with an overview of industry-specific AI solutions, illustrated by a concrete case developed with Takeda in life sciences.
