From BTP to SAP Business AI Platform: what it means for manufacturing, assets and supply chain
For more than a decade, SAP BTP meant one thing to the people running SAP in plants, mills and yards: a toolbox. Integration for connecting machines and ERP. Build services for extending S/4HANA without touching the core. A place to write code, store data and occasionally add a chatbot. It was infrastructure, important but invisible to the business.
At SAP Sapphire in May 2026, that toolbox changed shape. SAP Business Technology Platform, SAP Business Data Cloud and SAP Business AI came together as a single, governed environment: SAP Business AI Platform. It can look like a rebrand. Underneath, it is the clearest signal yet that SAP no longer sees its platform as something you build on, but as something that acts on your behalf. For anyone responsible for SAP in a manufacturing or asset-intensive business, that difference is the whole story.
Why this is more than a new name
SAP has been open about what drove the change. Customer feedback showed that the pieces of the old platform did not really work together: agent building sat in one place, business data in another, and governance was separate again. An agent built with a general-purpose AI model on its own could produce much the same result as one built on BTP. The platform that was meant to set SAP apart was not giving customers much they could not get elsewhere.
SAP treated this as urgent, and the fix was architectural rather than cosmetic. The build layer (Joule Studio), the data layer (SAP Business Data Cloud) and a new governance layer (SAP AI Agent Hub) now work as one. Whichever large language model a company chooses, it inherits that company's real business context, processes and rules the moment it is deployed. SAP describes the result as the foundation for the Autonomous Enterprise.
Agentic AI: from advice to execution
The most useful way to understand the shift is through agentic AI: software that does not just answer questions, but plans, decides and acts inside a business process, with a person reviewing the outcome rather than every step.
SAP's first generation of AI, the Joule assistant, was conversational. You asked a question, got an answer, then went and did the work yourself. The Business AI Platform is built for the next stage:
- Joule Studio 2.0, rolling out from June 2026, is a no-code and pro-code agent builder, with twelve months of free design-time access.
- SAP AI Agent Hub governs every agent that gets built, recording what it touched, what data it saw and who approved its actions.
- Agent-to-agent interoperability through the open Agent2Agent (A2A) protocol lets agents built outside SAP, even on another vendor's framework, call SAP's own agents for ERP tasks, with every call kept auditable.
- A choice of foundation models from several leading AI providers now powers Joule agents across HR, procurement and supply chain.
This is the part that should energise, not just inform, anyone who owns SAP in an operations-heavy business. SAP is betting that the value now lies in autonomous execution of end-to-end processes, not in dashboards that tell you something is wrong.
The honest caveat: governance is the real project
None of this is automatic or risk-free, and SAP says so plainly. The fear at board level is agentic chaos: uncontrolled agents roaming across a company's most valuable data and sharing it with people who should not see it.
That is why governance is a first-class layer of the platform rather than an afterthought: the Agent Hub, the audit trail, and a "company memory" built on SAP's process-mining capabilities. It is also why some early adopters are choosing to use SAP's ready-made agents rather than build large numbers of their own. For a business with a lean SAP team and a shop floor that cannot tolerate uncontrolled automation, that trust has to be earned through visible governance, not marketing language.
What this means for SAP leaders, starting now
The practical change is in the mandate, not just the tooling. The question moves from "which modules do we license?" to "which end-to-end processes are we willing to let an agent run, and what must our governance look like before we do?" Three moves are worth starting this year:
- Make clean core and business context prerequisites. Agents are only as good as the context they inherit: clean master data, documented processes and a disciplined, clean core. Without it, an agent is just a generic model with access to your ERP.
- Pilot with the agents already shipping. Production planning, asset health and field service dispatch map directly onto shop-floor pain, and availability runs through the second and third quarters of 2026.
- Start the governance conversation early. Decide who can approve an agent to act, what gets logged, and how data residency and sovereignty requirements are met.
The bigger picture
Retiring the BTP name is a small thing. The change behind it, from a platform that waits for developers to build something to one that ships agents ready to help run the plant, the maintenance crew and the supply chain, is not. For businesses that lean into it deliberately, with governance built in from day one, this is the most significant shift in how SAP technology touches operations since the move to S/4HANA itself.
If you are planning your first agents, see SAP Business AI Platform extensions and AI agents or read our explainer on SAP Business AI Platform and SAP BTP.
