SAP has outlined a framework for enterprise AI governance, arguing that it is essential for protecting profit margins as organizations move beyond statistical guesses to deterministic control. The company presented its position ahead of the AI & Big Data Expo North America, where governance of autonomous AI systems will be a central topic.

Manos Raptopoulos, Global President of Customer Success for Europe, APAC, Middle East and Africa at SAP, stated that the gap between near-perfect and perfect accuracy is not incremental but existential in enterprise environments. He noted that consumer-grade models often fail by ten percent when asked to count words in a document, illustrating the precision required for business-critical operations.

According to Raptopoulos, evaluation criteria for large language models deployed in production have shifted toward precision, governance, scalability, and tangible business impact. He identified the transition from passive tools to active digital actors as the primary governance moment facing corporate boards today.

Agentic AI Systems and Operational Risk

Agentic AI systems now possess the capability to plan, reason, orchestrate with other agents, and execute workflows autonomously. Because these systems interact directly with sensitive data and influence decisions at scale, Raptopoulos argued that failing to govern them exactly as one governs a human workforce exposes organizations to severe operational risk.

He warned that agent sprawl could mirror the shadow IT crises of the past decade, though with higher stakes. His framework requires agent lifecycle management, definition of autonomy boundaries, enforcement of policy, and continuous performance monitoring.

Integrating modern vector databases with legacy relational architectures demands substantial engineering capital, Raptopoulos said. Teams must restrict the agent’s inference loop to prevent hallucinations from corrupting financial or supply chain execution paths. These parameters drive up computational latency and hyperscaler compute costs, altering profit and loss projections.

Raptopoulos emphasized that governance becomes a hard engineering constraint rather than a compliance checklist. He stated that corporate boards must resolve three baseline issues before deploying agentic models: identifying accountability for an agent’s error, establishing audit trails for machine decisions, and defining thresholds for human escalation.

Geopolitical and Regulatory Challenges

Geopolitical fragmentation complicates these questions, Raptopoulos noted. Sovereign cloud infrastructures, AI models, and data localization mandates are regulatory realities in major markets including New York, Frankfurt, Riyadh, and Singapore. Enterprises must embed deterministic control directly into probabilistic intelligence, which he described as a C-suite mandate rather than an IT project.

Structuring relational intelligence for commercial operations remains dependent on data quality. Raptopoulos called this the data foundation moment. Fragmented master data, siloed business systems, and over-customized ERP environments introduce unpredictability at critical moments. If an autonomous agent relies on fragmented foundations to provide a recommendation affecting cash flow, customer relations, or compliance positions, operational damage scales instantly.

Data Foundation and Enterprise Intelligence

Raptopoulos argued that extracting tangible enterprise value requires advancing beyond generic large language models trained on internet-scale text. True enterprise intelligence must be grounded in proprietary corporate data, including orders, invoices, supply chain records, and financial postings embedded directly into business processes. He stated that relational foundation models optimized for structured business data will outperform generic models in forecasting, anomaly detection, and operational optimization.

The operational friction of making an over-customized ERP environment intelligible to a foundation model halts many deployments, he said. Data engineering teams spend excessive cycles sanitizing fragmented master data to create a baseline for the AI to ingest. When a relational model must interpret complex, proprietary supply chain records alongside raw invoice data, the underlying data pipelines must operate with zero latency. If data ingest fails, predictive capabilities degrade instantly.

Integrating legacy architecture with modern relational AI requires overhauling entrenched data pipelines. Engineering teams face indexing decades of poorly classified planning data so that embedding models can generate accurate vector representations. Raptopoulos advised boards to evaluate whether their current data estate is genuinely prepared rather than layering probabilistic intelligence over disjointed foundations.

Employee Interaction and Adoption

Enterprise application interaction is transitioning from static interfaces to generative user experiences, a development Raptopoulos flagged as the employee interaction moment. Instead of manually navigating complex software ecosystems, employees will express their intent to the system. He offered the example of a user instructing the software to prepare a briefing for their highest-revenue customer visit that week. The AI agents would then orchestrate workflows, assemble context, and surface recommended actions.

Raptopoulos stressed that workforce adoption remains conditional upon trust. Employees will only embrace these digital teammates if the systems are reliably governed and transparent in their decision-making processes.

SAP is expected to present further details on its enterprise AI governance framework at the AI & Big Data Expo North America. Industry observers anticipate that the discussion will focus on practical implementation steps for boards and engineering teams navigating the shift toward autonomous AI systems in regulated environments.