Enterprise AI as a Context Problem
Enterprise AI context describes how artificial intelligence systems understand, connect, and govern business data, processes, and permissions so models can take reliable actions inside mission-critical applications. Rather than chasing the biggest large language model or the flashiest autonomous agent, SAP argues that the real bottleneck in enterprise AI is context: knowing which customer, order, contract, or supplier an AI agent is dealing with, under which rules. At SAP Sapphire in Orlando, the company framed AI success as an integration and governance challenge, not a model horsepower race. According to SAP CTO Philipp Herzig, “What’s not differentiating is the LLMs… Use OpenAI models, use Anthropic models, whatever you like.” The value, in this view, lies in the surrounding infrastructure that connects models to ERP workflows, master data, and compliance constraints so they can handle real business work instead of demo-friendly chat.
Inside SAP’s Business AI Platform Strategy
SAP’s Business AI Platform brings together its Business Technology Platform, Business Data Cloud, Autonomous Suite, and Business AI services into a single “context layer” for enterprise AI. The platform’s goal is to give AI agents consistent access to business processes, data models, authorizations, and customer-specific extensions already embedded in SAP landscapes. That approach treats the model as a replaceable component and the platform as the real asset. SAP’s Autonomous Suite will ship with more than 50 domain-specific Joule Assistants orchestrating over 200 specialized agents across finance, supply chain, procurement, HR, and customer experience. Joule Work turns those assistants into a front door for workflows across SAP and non-SAP systems. SAP’s chief AI strategy officer Sean Kask states that the company will not build its own general LLM, preferring to partner with providers like Anthropic and Mistral while concentrating on domain models and context-rich orchestration.
Data Governance and the Supply Chain Context Advantage
A core claim in SAP’s SAP artificial intelligence story is that data governance AI, not model size, will separate winners from also-rans. The company is turning its ERP heritage into a structured context layer that spans relational data, master data, and process knowledge. Recent moves support this: SAP plans to acquire Dremio to make Business Data Cloud an Apache Iceberg-native lakehouse, has announced plans to acquire Prior Labs for tabular foundation models, and completed its acquisition of Reltio for master data management. For customers, this means AI agents can tap clean, governed supply chain and finance data rather than brittle exports and one-off integrations. SAP’s relational pretrained transformer, SAP-RPT-1.5, targets structured tables so teams do not need a new predictive model for every narrow task. In this model, tabular data stops being a back-office artifact and becomes the primary fuel for enterprise AI context.
Openness, Agents, and the New Governance Layer
SAP’s Business AI Platform also promotes openness for developers while keeping tight control over context and governance. Agents are built on open-source frameworks such as AutoGen and LangChain, and SAP says it will switch frameworks as better options appear. Developers can work in Joule Studio or bring tools like n8n for visual orchestration and Vercel for React front ends; SAP describes this as going “where the developers are.” To keep this variety in check, SAP is introducing AI Agent Hub, a service that discovers, inventories, and governs AI agents, LLMs, and MCP servers across SAP and non-SAP systems. SAP positions Agent Hub as a way to prevent a future where every department launches its own unmanaged agent. The company is even experimenting with “agent mining,” an extension of process mining that tracks what agents did, where they stalled, and whether they behaved as expected.
