Skip to content
Artificial intelligence. Clearly. RSS feed
AI NEWS AI

News. Evidence. Perspective.

Intel and SAP connect warehouse orders to autonomous robot tasks

A proof of concept in Chengdu links business requests, local AI and reported outcomes for material handling.

SAP described a warehouse proof of concept at Intel’s Chengdu site in China on 27 October. Transport tasks from SAP Extended Warehouse Management become autonomous robot missions, with outcomes fed back into the business system. SAP reports reduced transport distance and repetitive manual handling, without giving a numerical comparison.[1]

Intel’s account explains the division of work. SAP supplies business context and process rules; Intel’s hardware and software support local perception, navigation and action. The company describes combining CPUs, graphics processors and neural-processing units for these edge workloads. Status and exceptions return to enterprise systems. Intel presents the collaboration as a step toward a repeatable design across sites and robot suppliers.[2]

An instruction carries more than a destination

SAP’s warehouse reference architecture separates three abilities: carrying out physical operations, interpreting the business meaning of a task, and finding the correct storage location. Order priority and handling requirements belong to that middle layer. A robot needs the warehouse layout to find a bin; the associated business information helps establish what should move there and how it should be handled.[3]

Consider an illustrative task to move a container. Its destination answers where it belongs. Its priority helps determine when to do the work, while a handling requirement constrains the physical action. SAP’s architecture treats these as connected inputs to execution.[3]

The return message is part of the job

There is established integration machinery behind warehouse robotics. SAP’s 2023 guide describes a controller that receives warehouse orders and calls an adapter to create robot tasks. The adapter passes work to the physical robot, tracks its status and reports changes back. The guide also describes keeping orders locally so robots can continue processing during intermittent network connections.[4]

That older documentation provides useful context for what an enterprise connection must do. Dispatch, progress and completion are different states of the same job. Returning the state gives the receiving system a way to distinguish work allocated from work finished, and to base its next decision on what has happened on the floor.[4]

Follow the evidence

Sources & further reading

  1. Warehouse Pick & Place with Physical AI Robotics SAP Architecture Center · Accessed 28 Oct 2026
  2. Integration Services for SAP Warehouse Robotics SAP · 19 Aug 2023 · Accessed 28 Oct 2026