Christian Klein predicts that AI agents will overcome current accuracy hurdles (e.g., from 93% to 100%) for specific business functions like financial closing in a matter of months, not years. This will dramatically accelerate enterprise adoption.
Christian Klein predicts that AI agents are set to eliminate entire categories of work, specifically the controls and compliance checks that are currently full-time jobs for many in finance and HR. This shifts human focus to higher-value tasks.
Christian Klein argues that the "SaaS-pocalypse" narrative, where AI allows easy replication of complex software, is flawed. The real competitive moat for enterprise systems like ERP isn't the code, but decades of embedded process knowledge, data correlations, and regulatory compliance.
SAP's strategy against AI disruption isn't to build the best LLM, but to provide a unique "AI foundation." This layer adds business process knowledge, data context, and governance to any underlying model, making the generic AI truly enterprise-ready and creating a new defensible moat.
Christian Klein described reaching a tipping point where runaway AI usage forced SAP to institute formal token limits for employees based on their roles. This signals a new phase of enterprise AI where cost governance is becoming a standard operational control.
Christian Klein argues that reskilling employees—like moving on-premise developers to the cloud—is primarily a challenge of mindset. The desire to enter a "new world" of working is more crucial than the technical training itself for a successful transition.
SAP CEO Christian Klein reveals that his teams constantly test and switch the underlying AI models for their agents. This "model switching" has become a core strategy to manage runaway token costs, prioritizing the best price-to-outcome ratio over allegiance to a single frontier model.
Christian Klein points out that even advanced AI agents struggle to be accurate when fed data from disparate, messy systems. The bottleneck isn't the AI model itself but the underlying data quality and silos within a company.
According to SAP's CEO, the velocity of AI development and its impact on execution speed means traditional annual planning is obsolete. Companies must now adopt much shorter, faster planning cycles for everything from financials to workforce management to keep pace.
Christian Klein states that regulators in Brussels have realized that rules like the AI Act and Data Act have become too burdensome. He is confident that a rollback and simplification effort is underway, shifting the focus from regulating the technology itself to its societal impact.