# SAP adds governed gateway, n8n triggers and tabular AI models to Joule agents

published: 2026-10-27T08:05:29.066Z
topic: frameworks
source: https://siliconangle.com/2026/10/27/sap-advances-its-autonomous-enterprise-push-with-smarter-governed-joule-agents/
url: https://aiagentnews.io/stories/sap-adds-governed-gateway-n8n-triggers-and-tabular-ai-models-to-joule-agents

SAP announced new AI capabilities at its TechEd developer conference, expanding its Joule assistant and Business AI Platform. The updates center on governed agent access through Model Context Protocol and Agent2Agent, plus new tabular data models.

SAP announced the updates at its TechEd conference for developers. New Joule Work capabilities let agents interoperate with open protocols, specifically the Model Context Protocol and Agent2Agent, so Joule can gather context, make decisions and run multistep business workflows that can also draw on third-party data or services. A new orchestration layer lets other assistants and agents invoke Joule Agents through a governed gateway, and a recent update lets agents trigger n8n workflows through Joule Studio.

SAP also expanded SAP Business Data Cloud Ontology Manager, which defines relationships between company facts so agents can reason about how applications relate. The company said its acquisition of Reltio adds multidomain data management to that platform, and that its Dremio acquisition adds native Apache Iceberg and Apache Polaris support along with a federated query engine.

On the modeling side, SAP introduced SAP-RPT-2.0, a tabular model that adds anomaly and outlier detection, and SAP-RPT-TS for forecasting time-indexed metrics. SAP also said it is deploying SAP-RPT on the International Space Station in a partnership with Hewlett Packard Enterprise.

## Why it matters if you build agents

Developers building agents that call SAP systems may now route requests through a governed gateway rather than connecting directly, which changes how permissions are set up.

The n8n trigger is a deterministic option: an agent can decide when a workflow runs, while the workflow itself follows a fixed sequence of steps.

Teams working with structured business data get a tabular model that handles anomaly detection and forecasting without custom machine learning work.