SAP announced enhancements to its Joule AI assistant that enable AI agents to plan and execute multi‑step enterprise workflows, coordinating applications, data and agents across systems. The company emphasized governance and data control while introducing new tooling—such as Business Data Cloud, Knowledge Graph improvements and a central Enterprise AI Gateway—to help AI better understand company processes and to centralise oversight.
Practical change for workflows
SAP illustrated the technology with a common business scenario: resolving a mismatch between a supplier invoice and the purchase order. Today this often requires several employees to find documents, verify data, consult stakeholders and obtain approvals. The new AI agents can automate large parts of such processes while leaving key decisions under human review.
On the Joule Work interface, users can describe their goal in natural language and the system will coordinate the required applications, data and AI agents to achieve it.
Cross‑system collaboration and customization
SAP said the developments are not limited to SAP applications: AI agents from third‑party vendors can participate in the same workflow, provided appropriate access and security controls are in place. SAP will ship preconfigured AI assistants and agents across five major enterprise functional areas with industry‑specific customization options, and users can also create their own agents.
Data understanding and Knowledge Graph
A major enterprise AI challenge is that information is distributed across finance, procurement, manufacturing and sales systems, often representing the same business event in different ways. To address this, SAP introduced updates to Business Data Cloud and its Knowledge Graph to help AI better understand an organisation’s processes, data and operational rules. For example, when investigating a stock shortage, the AI can consider not only warehouse quantities but also expected orders, suppliers and production plans.
Data governance and control
SAP stressed that customers retain control over their data: the company will not hand over customer data to third parties for training or improving large language models. With AI taking on more autonomous tasks, governance becomes critical. To that end SAP announced the Enterprise AI Gateway, a central control plane for diverse enterprise AI systems that can standardise access policies, track AI usage and monitor costs. The gateway is planned for rollout in the first quarter of 2027.
Models and special use cases
SAP also presented new AI models that can support forecasting. A notable application is aboard the International Space Station: SAP and Hewlett Packard Enterprise (HPE) said they plan to run the SAP‑RPT model on the ISS’s HPE Spaceborne Computer‑2. On‑site analysis can aid in spotting technical anomalies, predicting failures, estimating missing sensor data and producing status reports without sending data back to Earth for processing.
When and where
These announcements were made at SAP TechEd 2026, and SAP expects to deploy the Enterprise AI Gateway in Q1 2027.
Conclusion
SAP’s new AI agents aim to automate multi‑step enterprise processes while improving the AI’s understanding of company data and providing tools for centralised governance and control. The company reiterated customer data ownership and demonstrated collaboration scenarios extending even to the International Space Station.
