Analyst(s): Brad Shimmin
Publication Date: May 14, 2026
SAP Knowledge Graph and SAP-RPT-1.5 use a unified semantic map and Retrieval-Augmented Prediction (RAP) to drive autonomous, production-ready business actions. The SAP Autonomous Suite leverages this metadata grounding and secure agent execution to transform traditional ERP workflows into high-precision, result-oriented outcomes.
What is Covered in This Article:
- The introduction of SAP Knowledge Graph as the foundational context layer for the autonomous enterprise.
- Advancements in tabular reasoning and precision-based AI with the launch of SAP-RPT-1.5 and Retrieval-Augmented Prediction (RAP).
- The structural expansion of SAP Business Data Cloud through native SAP HANA Cloud integration and the intent to acquire Reltio to unify master data.
- Strategic partnerships with Palantir, AWS, and NVIDIA to accelerate data migration, zero-copy integration, and secure agent execution.
The News: At SAP Sapphire 2026 in Orlando, SAP unveiled a strategic transition toward the autonomous enterprise. This vision moves past simple conversational assistants toward a unified SAP Business AI Platform where AI agents execute end-to-end business processes. The event centered on the realization that for AI to be mission-critical, it must move beyond the 80% accuracy of generic models—a point emphasized by CEO Christian Klein’s anecdote regarding an AI-generated unicorn with three ears—and anchor itself in the specific, structured reality of enterprise data.
The primary move was the unification of the SAP Business Technology Platform (BTP), SAP Business Data Cloud (BDC), and SAP Business AI into a single governed environment. Technically, the most significant milestone is the SAP Knowledge Graph. This solution provides a structured map of 7.3 million data fields, business entities, processes, and relationships across the SAP landscape. This graph allows AI agents to navigate complex organizational structures without losing sight of the underlying business logic, identity rules, or security permissions.
Beyond context, SAP focused on execution through the SAP Autonomous Suite and Joule Work. The company announced 224 specialized agents and 51 Joule Assistants designed to handle high-volume, time-sensitive tasks across finance, supply chain, procurement, HCM, and customer experience. Supporting this is a deeper open data ecosystem, highlighted by a bi-directional zero-copy integration between SAP Business Data Cloud and Amazon Athena, and a partnership with Palantir to leverage AIP for complex data migrations. These moves suggest a pragmatic focus on building a substrate for software that reasons and acts based on real-world constraints.
Precision Over Prose: Why SAP Knowledge Graph is the Secret to Production-Ready AI
Analyst Take: The success of generative AI in the enterprise has long been stymied by a lack of grounding. While large language models excel at prose, they are notoriously poor at understanding the intricate dependencies of a global supply chain or a complex financial close. According to Futurum Research, enterprise leaders increasingly identify the lack of business-specific metadata as the primary hurdle preventing AI from reaching production-readiness. With the launch of the SAP Knowledge Graph, SAP is addressing this metadata drought head-on.
By mapping the relationships between processes and data entities, SAP provides the connective tissue that agents need to operate reliably. This architecture provides a high-fidelity map of the business that complements existing large language models. This structural grounding is bolstered by SAP-RPT-1.5, a specialized tabular AI model in the SAP Foundation Model family. While generic models struggle with structured database logic, SAP-RPT-1.5 delivers on the need for precision by offering Retrieval-Augmented Prediction (RAP) and column-level explainability. These features allow users to identify exactly which business attributes influenced a prediction, which is essential for establishing trust in automated decision-making.
The Technical Necessity of the Knowledge Graph
The SAP Knowledge Graph represents a departure from the traditional retrieval-augmented generation (RAG) patterns that dominated the early stages of generative AI. Standard RAG relies on vector similarity, which works well for unstructured text but fails when applied to the relational complexities of an ERP system. An agent needs to know that a change in a bill of materials in a German manufacturing plant has specific tax implications and downstream logistics consequences. Traditional vector databases struggle to maintain these logical hierarchies.
By codifying 7.3 million data fields and their interrelationships, SAP is essentially providing its agents with the equivalent of a senior consultant’s tribal knowledge. This graph enables agents to perform multi-hop reasoning—understanding that entity A is connected to process B, which is governed by policy C. This level of semantic awareness allows for a much lower fault tolerance in data retrieval, moving the needle closer to the 100% accuracy required for financial and operational workflows.
Expanding the Data Fabric through Reltio and SAP HANA Cloud
The integration of SAP HANA Cloud and the intent to acquire Reltio suggest a rigorous commitment to building a robust business data fabric. Futurum Research indicates that a growing number of Global 2000 organizations are now piloting agentic workflows that connect autonomous reasoning to transactional systems. However, these workflows fall apart if the master data is fragmented. Bringing Reltio’s multi-domain master data management into the SAP Business Data Cloud ensures that Joule agents are reasoning over a single, golden record of truth across SAP and non-SAP sources.
MDM has historically been a back-office optimization task. In the agentic era, it becomes a prerequisite for survival. If an agent is tasked with optimizing “spend” but cannot reconcile “supplier X” in the procurement system with “vendor Y” in the finance system, the resulting “autonomous” action will be flawed. SAP is effectively positioning Master Data Governance as the foundational layer of its AI stack, ensuring that Joule and its assistants operate on high-quality, connected data.
From SaaS to Software-as-a-Result
Furthermore, the transition from Software-as-a-Service to what SAP calls Software-as-a-Result marks a fundamental change in user engagement. Joule Work redefines the user interface by moving toward a dynamic workspace that adapts to user intent. Instead of navigating through fragmented, transactional screens, users describe a desired outcome, and Joule orchestrates the necessary agents and data to achieve it.
This adaptive experience, supported by NVIDIA OpenShell for secure runtime execution, moves the enterprise toward a “headless” or “app-less” paradigm where the software generates the necessary capabilities on the fly. This is a bold gamble on the future of the human-computer interface. For decades, ERP was synonymous with complex, static menus. SAP is now betting that the future of enterprise software is a blank prompt backed by a massive, semantic understanding of the business.
The Role of Partnerships in Solving the Migration Problem
One of the most pragmatic elements of the Sapphire announcements was the admission that AI cannot solve the migration problem alone. The partnership with Palantir to use AIP for data-migration scenarios and the zero-copy integration with Amazon Athena show that SAP is embracing an open data ecosystem. The ability to query data in Amazon Athena from the SAP Business Data Cloud without moving it reduces both latency and egress costs—two massive implementation challenges for any modern data architecture.
The agent-led transformation tools, which SAP claims can reduce ERP migration efforts by 35%, target the “clean core” problem. By using agents to automate system analysis and code remediation, SAP aims to remove the technical debt that has kept many customers on-premises for years. If these tools deliver on their promise, the path to the autonomous enterprise will look much more predictable for the average enterprise executive.
What to Watch:
- Keep a close eye on the adoption of the Agent-to-Agent (A2A) and Model Context Protocol (MCP) standards. The ability for Joule to securely call on third-party agents from Microsoft or Google using the SAP Knowledge Graph context will determine if SAP becomes the central nervous system of the enterprise.
- Monitor how quickly Reltio’s capabilities are natively embedded into the SAP Business Data Cloud capacity pool. A seamless transition will be critical for customers struggling with data quality in hybrid environments, especially as SAP aims to merge SAP and non-SAP data products.
- Watch the performance of SAP-RPT-1.5 in high-stakes scenarios like the Autonomous Close Assistant. If it can reliably compress weeks of manual reconciliation into days by leveraging the Knowledge Graph, it will set a new benchmark for ROI in autonomous finance.
- The partnership with Palantir and the new agent-led migration tools aim to reduce ERP transition efforts by 35%. Whether this actually accelerates the move from on-premises ECC to Cloud ERP remains the ultimate test of SAP’s data strategy and its ability to remove technical debt.
You can read the full press release outlining the many announcements made at Sapphire 2026 at SAP’s website.
Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum as a whole.
Other Insights From Futurum:
Can SAP Finally Kill the ETL Monster? The Dremio and Prior Labs Acquisition Explained
Tableau Dismantles the BI Dashboard With a Graph-Powered Leap Into Headless, Agentic Analytics
Author Information
Brad Shimmin is Vice President and Practice Lead, Data Intelligence, Analytics, & Infrastructure at Futurum. He provides strategic direction and market analysis to help organizations maximize their investments in data and analytics. Currently, Brad is focused on helping companies establish an AI-first data strategy.
With over 30 years of experience in enterprise IT and emerging technologies, Brad is a distinguished thought leader specializing in data, analytics, artificial intelligence, and enterprise software development. Consulting with Fortune 100 vendors, Brad specializes in industry thought leadership, worldwide market analysis, client development, and strategic advisory services.
Brad earned his Bachelor of Arts from Utah State University, where he graduated Magna Cum Laude. Brad lives in Longmeadow, MA, with his beautiful wife and far too many LEGO sets.
