SAP executives say companies deploying AI agents across enterprise workflows need to treat AI as an integrated platform rather than a collection of experiments, according to remarks published November 16.
Thorsten Leiduck, SVP and global head of SAP Business AI Platform Ecosystem, told MIT Technology Review that reliable AI deployment depends on “a connected, semantically rich layer that reflects their business.” He said accuracy requirements shift as AI moves into critical processes: while 80% accuracy may suffice for some everyday applications, that level “is a very terrible result” in areas such as supply chains or financial close.
Satinder Pal Singh, associate VP North America for SAP BAIP and SAP AI Strategy at Infosys, recommended evaluating AI agent use cases through impact, feasibility and scalability. “A pilot that can’t scale is just an experiment,” he said, adding that companies should identify where agents can move business metrics, assess whether their data and talent support the use case, and determine if successful applications can be replicated.
Singh said governance becomes more important at scale. “The absence of governance doesn’t speed up AI adoption,” he said. “It creates a risk backlog” that can force disruptive audits. He described the longer-term opportunity as a shift from AI as a productivity tool to “AI as a decision-making partner.”
Both executives anticipate agentic systems coordinating work across applications, with customer-specific knowledge graphs and governed data becoming critical enterprise assets.