SAP has announced a broad set of AI updates at SAP TechEd, expanding its tools for building enterprise agents, connecting business data and governing automated workflows. The changes span Joule Studio, Joule Work and SAP Business AI Platform, alongside new capabilities for SAP Business Data Cloud and the company’s tabular foundation models. SAP is positioning the portfolio as infrastructure for an “autonomous enterprise,” where AI agents can reason over business context and execute multistep processes across SAP and third-party systems, subject to enterprise controls.
SAP Pushes Enterprise AI From Assistance to Execution
Enterprise AI adoption is moving beyond chatbots and standalone copilots toward agents that can coordinate tasks across business applications. SAP’s latest announcements target that shift by combining agent development tools, business-data context, interoperability and centralised governance in one platform strategy.
At the centre is SAP Business AI Platform, which SAP describes as a common environment for building, contextualising and governing AI across SAP and third-party systems. Joule Studio provides no-code and pro-code tools for developing and extending agents, with access to more than 70 AI models. Customers in its early adopter programme can also design and run n8n workflows within Joule Studio, with broader availability planned later.
Joule Work is the user-facing experience for expressing an intent and having AI assist with the work. New agentic reasoning capabilities are designed to plan and orchestrate multistep workflows across SAP environments. SAP is also expanding interoperability through Agent2Agent (A2A) and Model Context Protocol (MCP), alongside a governed gateway that lets third-party agents securely invoke Joule Agents.
The strategic difference is execution. Instead of stopping at a generated answer, an agent could coordinate a sequence of actions across business systems. The value of that approach depends on how accurately the agent interprets context, handles exceptions and operates within the permissions assigned to it.
Business Context Becomes a Core AI Layer
A persistent challenge for enterprise AI is that general-purpose models do not automatically understand a company’s internal processes, data relationships or business rules. An agent may generate a plausible response but still lack the context needed to make a dependable operational decision.
SAP is addressing this through SAP Business Data Cloud and SAP Knowledge Graph. New Knowledge Graph capabilities will incorporate additional SAP and non-SAP sources, allowing users to query customer-specific information in natural language. The SAP Business Data Cloud Ontology Manager will add an AI-native environment for defining relationships across enterprise data, helping agents interpret how information from different systems and departments fits together.
SAP is also extending its data ecosystem through expanded integrations with Dremio, Databricks and Reltio. The Dremio capabilities include support for Apache Iceberg and Apache Polaris, along with federated queries designed to provide access to data without requiring it to be moved or copied. SAP’s expanded Databricks integration, branded SAP Databricks, brings data engineering and AI capabilities into Business Data Cloud. The Reltio solution is intended to unify customer, supplier and product records across SAP and non-SAP environments.
These additions matter because agent performance depends not only on the model, but also on data quality, access and the relationships used to interpret information. Shared business context can reduce the need to build separate integrations and definitions for every AI application. Actual improvements in accuracy, however, will depend on implementation quality and the completeness of the underlying data.
Governance Moves Into the Agent Infrastructure
As AI agents gain the ability to interact with enterprise systems, organisations need controls that extend beyond model selection. SAP announced a new Enterprise AI Gateway to centralise authentication, policy enforcement, routing, monitoring, usage tracking and auditing of AI interactions.
The company is also expanding governance capabilities through Joule Assistant and SAP AI Agent Hub, alongside an AI Governance Assistant designed to help customers address EU AI Act compliance requirements. Joule Studio is gaining observability and administrative controls to help developers and administrators monitor agent behaviour, model usage and spending.
These controls address practical barriers to enterprise adoption. Organisations need to know which agents can access specific information, what actions they can execute, how much they cost and whether their activity can be audited. A governed agent gateway could also make it easier to coordinate agents from different providers without giving every system unrestricted access.
The presence of governance features does not automatically make an AI deployment compliant. Companies will still need to classify their use cases, assess risk, configure permissions and establish oversight processes appropriate to their operations.
SAP says selected customers can deploy and run custom agents on its managed Joule Studio runtime free through December 2026 under a time-bound promotion. Design-time tools for building and testing agents and n8n workflows with AI assistance will remain free within fair-use limits.
SAP Expands Tabular AI and Forecasting
The announcements also extend SAP’s work on models designed for structured business data. Its latest tabular foundation model, SAP-RPT-2.0, is intended to expand beyond classification and regression into anomaly and outlier detection and embedding-based analysis. SAP-RPT-TS is designed to predict values across time-indexed business metrics, including forecasting scenarios associated with SAP Integrated Business Planning.
Structured enterprise data presents different challenges from the text used by many generative AI systems. Financial figures, supply-chain records and operational measurements often sit in related tables, where useful analysis depends on relationships between fields and time periods. Models built for these tasks could help reduce the manual feature engineering often required by conventional machine-learning projects.
SAP also announced plans to run SAP-RPT on the HPE Spaceborne Computer-2 aboard the International Space Station. The proposed use cases include anomaly triage, predicting potential failures, reconstructing missing sensor values and generating mission-health reports without sending the data back to Earth for analysis. The announcement describes a planned deployment, rather than reporting completed in-orbit validation.
Partners Extend the Autonomous Enterprise Strategy
SAP is expanding its ecosystem through a collaboration with Intel focused on connecting enterprise workflows with autonomous robots, bridging information technology and operational technology through physical AI.
New SAP solution extensions with Docusign, Parloa and Redwood target agreement management, conversational agent capabilities and managed file transfer. These integrations reflect a wider strategy to extend automation across workflows that traditionally cross application boundaries and require human handoffs.
The broader challenge for SAP is to turn a large portfolio of agents, data services, models and integrations into a coherent operating environment. Customers will judge the platform on whether it reduces implementation complexity, maintains reliable controls and delivers measurable business outcomes.
SAP’s TechEd announcements outline a shift from AI that helps employees find information to systems designed to take action within enterprise processes. Its advantage will depend on how effectively the company combines its business-process knowledge with open interoperability, trusted data and governance. For customers, the central question is no longer just what an AI agent can generate, but what work it can safely complete.
Market Landscape
Enterprise AI is moving toward agent-based systems that combine model choice, company-specific data, workflow execution and governance. SAP’s strategy is to embed those capabilities into its existing enterprise software ecosystem while allowing agents and data services from third parties to participate.
- Agentic enterprise software: Joule Work and Joule Studio target multistep workflows rather than isolated AI-generated responses.
- Business-context grounding: SAP Knowledge Graph and Business Data Cloud aim to make enterprise data relationships available to AI agents.
- Interoperability: A2A, MCP and a governed agent gateway are designed to support collaboration across different AI systems.
- AI governance: Centralised access controls, observability, usage tracking and audit records address risks associated with agents taking actions.
- Specialised AI models: SAP-RPT-2.0 and SAP-RPT-TS target structured-data analysis and forecasting beyond general-purpose language model use cases.
Top Insights
- SAP is expanding Joule Work, Joule Studio and SAP Business AI Platform to support AI agents that coordinate multistep enterprise workflows.
- SAP Business Data Cloud Ontology Manager and Knowledge Graph are designed to ground agent decisions in relationships across SAP and third-party data.
- A new Enterprise AI Gateway will centralise authentication, policy enforcement, routing, monitoring, usage tracking and auditing of AI interactions.
- SAP-RPT-2.0 expands structured-data analysis, while SAP-RPT-TS targets forecasting across time-indexed business metrics.
- SAP’s planned deployment of SAP-RPT on the International Space Station would test enterprise tabular AI in an edge-computing environment.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
