Enterprise AI initiatives rarely fail because the models lack power. They fail because those models lack the business context required to make reliable decisions inside real processes.
SAP CTO Philipp Herzig has consistently linked AI value to practical enterprise constraints such as integration complexity, compliance, AI sovereignty and data quality, the factors that determine whether an agent can be trusted to act. This is especially relevant for organisations already running SAP, where AI must understand stock availability, approval limits, supplier risk, invoice exceptions and customer commitments exactly as configured. In enterprise AI, context is both the bottleneck and the opportunity.
SAP’s response is SAP Business AI Platform, launched at SAP Sapphire Orlando in May 2026. It reorganises the SAP technology stack around a single thesis: business context, not model capability alone, is what moves AI from answering questions to executing work.
For RISE with SAP and SAP S/4HANA Cloud, Private Edition customers, this article explains the platform, the components required for autonomous execution, and a practical path to getting started. It begins by demonstrating how the building blocks fit together.
Inside SAP Business AI Platform: The Architecture for the Autonomous Enterprise
SAP Business AI Platform brings SAP Business Technology Platform, SAP Business Data Cloud and SAP Business AI into a single governed enterprise AI foundation, built on the SAP S/4HANA system RISE customers already run.
SAP BTP provides the runtime, integration, extension and identity layer that connects AI back into enterprise systems and is typically the starting point because most RISE customers are already entitled to SAP BTP. Additionally, SAP Business Data Cloud unifies SAP and non-SAP data, ensuring agents work from current operational information. AI Foundation supplies the models, while SAP Knowledge Graph adds the business context that grounds AI in SAP-specific processes and authorisation context.
Joule Studio is where teams design agents that use this context and trigger business actions. SAP Integration Suite connects them across SAP, third-party and partner systems, and SAP AI Agent Hub governs agents, models and MCP servers with enterprise architecture context from SAP LeanIX.
Together, these blocks form one operating model for the Autonomous Enterprise and define where SAP draws the line between an assistant that helps and an agent that acts.
From Assistant to Agent: Joule and the Autonomous Suite
SAP distinguishes Joule Assistants, which answer questions and support tasks, from Joule Agents, which execute defined workflows with greater autonomy. At Sapphire 2026, SAP positioned Joule and the Autonomous Suite with more than 50 domain-specific Assistants orchestrating over 200 agents across Finance, Spend, Supply Chain, HCM and Customer Experience. Many operate within SAP S/4HANA processes that RISE customers already use, demonstrating SAP’s move from assistance to execution without re-platforming.
This is where embedded AI differs from bolt-on AI: SAP agents sit closer to application logic, business context and authorisation models, so an invoice agent can use the three-way match because the process already exists in the system. With that depth, the remaining task is to define clear execution boundaries (when to recommend, draft, act or hand to a human approver), and for RISE customers, that starting point is closer than it looks because most of the foundation is already in place.
From RISE to Autonomous: A Practical Path for SAP Customers
RISE with SAP and SAP S/4HANA Cloud, Private Edition customers are already close to the Autonomous Enterprise. RISE provides the clean-core operating model and managed cloud foundation that SAP Business AI Platform extends. The practical question is which AI components to add and in what sequence.
A RISE subscription already includes the key building blocks: SAP S/4HANA as the system of record and, in most cases, SAP BTP for runtime, integration and extension. Trusted data, context, models, agent tooling and governance are added on top of the landscape already in operation.
At a glance, this is what a typical RISE customer already owns and what they need to add for autonomous AI.
| Already part of your RISE landscape | What must be added for autonomous AI |
| System of record and master data: SAP S/4HANA Cloud, Private Edition | Trusted data foundation: SAP Business Data Cloud |
| Runtime, integration and extension: SAP Business Technology Platform (typically included) | Models and business context: AI Foundation + SAP Knowledge Graph |
| Clean-core operating model and managed cloud landscape | Agent experience and build environment: Joule + Joule Studio |
| Identity, security and access model | Connectivity and governance: SAP Integration Suite + SAP AI Agent Hub |
Here is what each to-be-added component does and why a RISE customer needs it.
| Component | What it does | Why RISE customers need it |
| RISE with SAP: SAP S/4HANA Cloud, Private Edition | The live system of record holding process and master data | Already owned – it provides the business context required for reliable agent execution |
| SAP Business Technology Platform | Runtime, integration, extension and identity layer for AI use cases | Typically included with RISE – the foundation on which every other component operates |
| SAP Business Data Cloud | Governed data foundation unifying SAP and non-SAP data | Provides agents with current, consistent operational data in place of stale extracts |
| AI Foundation | Model access, generative AI hub, SAP domain models and semantic services | The intelligence layer beneath Joule and SAP’s embedded AI scenarios |
| SAP Knowledge Graph | Encodes business meaning across tables, objects, rules and relationships | Enables agents to recognise that a purchase order, supplier, material, goods receipt and invoice form a single process |
| Joule and Joule Studio | The assistant/agent experience and the governed environment to build agents | The environment in which agents are designed to use the SAP context and trigger business actions |
| SAP Integration Suite | Connectivity across SAP, third-party and partner systems | Enterprise processes extend beyond a single SAP application; agents must operate across the wider landscape |
| SAP AI Agent Hub | Governance and control plane for agents, models and MCP servers | Provides the inventory, auditability and lifecycle control required for regulated processes |
Knowing what each component does is one thing. Knowing which ones are required on day one is another.
Mandatory or Optional: What You Need on Day One
| Component | Mandatory vs Optional
To get started |
Why |
| SAP S/4HANA Cloud, Private Edition (RISE) | Mandatory | Your system of record and the source of business context; already owned |
| SAP Business Technology Platform | Mandatory | Runtime, identity and integration foundation the AI stack runs on; usually included with RISE |
| AI Foundation | Mandatory | Model access, generative AI hub and SAP domain models beneath Joule |
| SAP Knowledge Graph | Mandatory (part of AI Foundation) | Business meaning that grounds agents in real process relationships |
| Joule and Joule Studio | Mandatory | The agent experience and the governed environment to build and run agents |
| SAP Business Data Cloud | Recommended
|
Not strictly required if the first use case relies on data already in SAP S/4HANA; becomes essential once agents draw on governed, cross-system or non-SAP data |
| SAP Integration Suite | Optional
|
Required when agents must reach non-SAP, third-party or partner systems |
| SAP AI Agent Hub | Optional, essential at scale | Governance, audit and lifecycle control for production and regulated processes |
Bringing those mandatory and optional components together makes the target architecture clear.
A Reference Architecture for RISE Customers
Each layer builds on the one beneath it: RISE provides the system of record at the base, while Joule delivers the autonomous experience at the top. A RISE customer already owns the bottom two layers; the layers above are what they need to add. Architects typically focus on three practical considerations next: clean core, target architecture, and agent governance.
For Architects: Landscape and Integration Considerations
Clean core is the prerequisite, and AI belongs in the target architecture from day one. SAP Business AI Platform models, including SAP-ABAP-1 for ABAP development and SAP-RPT-1 for process and tabular reasoning, work best when extensions, integrations and custom code are aligned with ABAP Cloud and Clean Core practices. A Clean Core assessment is therefore the right starting point. From there, define agent data access, user identity propagation and integration boundaries across SAP S/4HANA, SAP Business Data Cloud and non-SAP systems, and enforce them through SAP Integration Suite and BTP identity services.
Establish agent controls before scaling. SAP AI Agent Hub should set the approved data domains, action thresholds, human checkpoints, logging standards and retirement criteria from the outset. Begin with a single high-value, cross-functional use case where context-rich agents clearly outperform bolt-on AI. Then, translate those decisions into a three-phase delivery plan.
A Phased Path to Production
| Phase 1: Establish the foundation | Phase 2: Deploy your first agents | Phase 3: Scale and govern |
| Confirm the SAP BTP entitlement, enable SAP Business Data Cloud and SAP Knowledge Graph, and prepare your master data and Clean Core. | Select one high-value, cross-functional use case (for example, e-invoice exception handling), build it in Joule Studio and define human-in-the-loop checkpoints.
|
Onboard agents into SAP AI Agent Hub, formalise the agent control plane, and extend the model across Finance, Supply Chain, HR and Procurement. |
In principle, this is a clean three-phase plan. The next section demonstrates how it works in practice, end-to-end.
Agent in Action: End-to-End Across Your SAP Landscape
Components only matter when they work together. The walkthrough below follows a familiar process – e-invoice exception handling – from the moment an exception is raised to the point it is resolved. It shows what each part of the stack does at every stage and how the existing RISE systems fit together.
SAP S/4HANA → Business Data Cloud → Knowledge Graph → AI Foundation → → Joule + Joule Studio → Integration Suite → AI Agent Hub
| Step | What happens | SAP component in play |
| 1 | A supplier invoice fails the three-way match in the core system.
|
SAP S/4HANA Cloud, Private Edition |
| 2 | The agent gathers the invoice, purchase order, goods receipt and supplier master, including any non-SAP data. | SAP Business Data Cloud |
| 3 | The agent learns how the purchase order, goods receipt, invoice, tax code and tolerance rules relate to one another. | SAP Knowledge Graph |
| 4 | The exception is evaluated against tolerances and policy, and a resolution is drafted. | AI Foundation (with SAP-RPT-1) |
| 5 | A Joule agent, built in Joule Studio, runs the workflow and decides whether to recommend or to act. | Joule and Joule Studio |
| 6 | Approved actions are posted back to the core system and notify the supplier through the relevant channel. | SAP Integration Suite |
| 7 | Every step is logged, attributed and made auditable, and the agent is inventoried. | SAP AI Agent Hub |
| 8 | Above a set threshold, the agent routes the case to a human approver before anything is posted. | Joule (in Microsoft 365 or SAP Fiori) |
The goal is not any single component but the hand-offs between them. A bolt-on tool can read an invoice; it struggles to move cleanly from data to context to decision to action while staying inside your security and governance model. For a RISE customer, that flow runs on systems you already operate. That is what makes end-to-end autonomy realistic rather than aspirational because the context that makes it work is already inside your SAP estate.
Context Is the Moat for the Autonomous Enterprise
Models will continue to advance, but in mission-critical SAP processes, model quality is only part of the equation. The greater challenge is giving AI reliable access to the business context it needs to make trustworthy decisions: rules, authorisations, master data, exceptions, controls and the consequences of every action.
This is the strategic value of SAP Business AI Platform for RISE customers. It surrounds foundation models with trusted data, business context, connectivity and governance, advancing AI from an assistant that supports work to an agent that executes it. Since RISE already provides the system of record and, in most cases, SAP BTP, adopting SAP Business AI Platform is largely an extension of the existing landscape.
The path is deliberately practical: build on what is already owned, introduce data, context, models, agent tooling and governance in sequence, and demonstrate value on one high-value process before scaling. Models will continue to commoditise; the business context embedded in the SAP estate will not, and that is the durable advantage RISE customers can convert into autonomous execution.
The next step is more straightforward than it appears: begin where the business already operates, select one high-value process, define a clear agent boundary, and treat the existing SAP estate as the authoritative source of truth.
The context already exists within the enterprise – SAP Business AI Platform converts it into measurable business action.
Ready to start your Autonomous Enterprise journey?
Get in touch with our SAP experts and let’s explore your SAP Business AI roadmap together.
