SAP Unveils Physical AI, Extending the Autonomous Enterprise into the Real World

SAP has introduced Physical AI by SAP, a capability that links robots, drones and other physical systems directly to the business processes that run an organisation. The company says the approach turns physical execution into a governed, auditable extension of the enterprise, rather than a separate layer of automation that operates on its own.

Enterprises have invested heavily in both business AI and physical systems such as industrial robots, humanoids, sensors and inspection drones. Business AI delivers insight and recommendations on finance, inventory and supply chains, while physical systems operate across production lines, warehouses, field service teams and logistics networks. Yet the two rarely connect. Physical operations produce a constant stream of information about asset health, task completion, quality, throughput and exceptions, but much of it never reaches the business record. SAP identifies this gap as a common source of operational friction, contributing to delayed responses, poorly informed decisions and unrecorded physical work, which in turn erodes revenue and margins and raises risk exposure.

SAP’s answer is to make the system that runs the business the same one that directs physical operations. Enterprise knowledge held in SAP triggers physical actions. Physical signals arrive with business context already attached. Verified outcomes from the field flow back into SAP as structured data, and every action leaves a traceable record tied to the relevant business process. For organisations already running SAP, physical execution inherits the intelligence, context and governance already in place.

The capability is built on four principles. First, physical actions are grounded in business context, including order priorities, safety rules, inventory status and compliance constraints, so that each task has a clear reason and the business system stays in control of what is executed and when. Second, physical evidence feeds digital workflows, with each task creating a complete audit trail and observations returning to SAP automatically. Third, execution is governed through SAP Business AI Platform, with traceability, auditability and human-in-the-loop control, while physical data enters SAP Business Data Cloud as new structured data products that enrich planning and predictions. Fourth, the approach is vendor and hardware agnostic, so customers keep their choice of devices, physical systems and AI models, with SAP supplying context, orchestration and governance and partners supplying physical execution.

SAP highlighted three examples. At Arçelik-LG, a joint venture between Arçelik and LG Electronics that manufactures air conditioners, a humanoid robot executes and records inbound storage tasks using SAP Extended Warehouse Management. This addresses the costly problem of incorrect storage placements, with potential savings of millions of euros. At Intel, a proof of concept at the Chengdu site translates warehouse transport tasks created in SAP into autonomous robot missions, combining SAP’s workflow governance with Intel’s edge AI and robotics technology. Material movement is automated end to end, reducing transport distances and repetitive manual handling. At Holcim, quadruped robots inspect conveyor belts in dark, dusty areas and kilns in extreme temperatures, while drones inspect assets at heights that are hard for people to reach. A maintenance plan in SAP triggers each inspection, and a maintenance notification is created back in SAP once it is complete.

Key takeaways

• Physical AI by SAP connects robots and physical systems to SAP business processes, so physical actions are driven, governed and recorded by the enterprise system.
• It aims to close the gap between what a business knows and what physical operations learn on the ground.
• It rests on four principles: business-driven action, physical evidence in digital workflows, enterprise governance and vendor-agnostic execution.
• Early examples span warehouse logistics at Arçelik-LG and Intel and industrial asset inspection at Holcim.
• SAP positions the capability as an extension of its Autonomous Enterprise vision into the physical world.