SAP has taken artificial intelligence beyond the realm of language. At the SAP TechEd 2025 conference in Berlin this week, the company introduced a new class of AI designed to predict business outcomes rather than words. The result is what SAP claims to be the first enterprise relational foundation model, called SAP-RPT-1, or Relational Pre-trained Transformer.
It represents a shift in what a foundation model can be: rather than being a generator of sentences, it is a predictor of decisions.
SAP describes SAP-RPT-1 as an AI that “can make fast and accurate predictions for common business scenarios like delivery delays, payment risk, or sales order completion”. It uses the relationships between business data rather than linguistic patterns to anticipate what happens next in an organisation’s operations. In simple terms, it reads the business, as opposed to mere text.
To encourage developers to explore what that means in practice, SAP has launched a free playground environment where they can test predictive scenarios, simulate business situations, and see how relational AI behaves when exposed to live data. For developers accustomed to fine-tuning language models, it represents a different kind of creativity, rooted in the mechanics of business itself.
“SAP’s announcements today give developers the tools they need to deliver at the speed of AI,” says Muhammad Alam, member of the executive board of SAP. “Innovations across SAP’s unique flywheel of applications, data and AI put developers in the driver’s seat – where they belong.”
That flywheel was clearly visible across the rest of TechEd’s announcements.
The company’s SAP Build platform, its centrepiece for enterprise application development and automation, has been re-engineered to give developers more freedom to use tools they already rely on.
Developers who prefer agentic development environments like Cursor, Claude Code, Cline and Windsurf can now use SAP development frameworks through new SAP Build local Model Context Protocol Servers. Visual Studio Code users can access SAP Build capabilities directly in their existing development environment via a new SAP Build extension, which will later also appear in the Open VSX Registry for use with other development environments.
SAP revealed plans for integration with automation platform n8n, allowing its Joule Studio agents to work alongside n8n’s own agents. This kind of cross-agent collaboration reflects the company’s growing emphasis on orchestration: enabling multiple intelligent systems to coordinate tasks across applications and departments.
The data layer that feeds these systems is also evolving. Every intelligent application depends on trusted data, and SAP is extending its reach through SAP Business Data Cloud. A new SAP Snowflake solution extension brings Snowflake’s managed data and AI capabilities directly to SAP customers, giving them “the flexibility to choose the right compute and storage for each data and AI workload, while maintaining governance, interoperability and business context”.
The announcement was reinforced by a new SAP Business Data Cloud Connect partnership with Snowflake, adding to existing integrations with Databricks and Google Cloud. The result is a more open, federated data ecosystem that lets developers work with SAP data wherever it resides.
A new data product studio capability in SAP Business Data Cloud now allows developers to turn raw data into ready-to-use assets known as data products, designed for analytics, AI and application development. At the same time, the SAP HANA Cloud knowledge graph engine can automatically generate knowledge graphs. This means it maps relationships across SAP database tables, columns, and data models to reveal how data fits together. For developers, it is a new way of seeing how information connects across systems, turning structure into insight.
SAP is also extending its Joule AI portfolio, which now includes new assistants that can coordinate multiple agents across workflows, departments and applications. These assistants are designed to “plan, initiate, and complete complex tasks spanning finance, supply chain, HR, and beyond.”
Among the new offerings is an agent for business process analysis, helping teams understand how processes actually run, identify inefficiencies, and uncover opportunities to optimise workflows and achieve measurable improvements.
SAP pledged to equip 12-million people worldwide with AI-ready skills by 2030, through a partnership with Coursera that expands hands-on training and certification in practical AI tools. The goal, SAP says, is to make AI accessible to “people everywhere”.
* Arthur Goldstuck is CEO of World Wide Worx, editor-in-chief of Gadget.co.za, and author of The Hitchhiker’s Guide to AI – The African Edge.