“Trained in digital twins and powered by physical AI, humanoid robots can reduce worker injuries... and lower overtime costs and dependency on temporary labor,” said Christian Souche, advanced robotics lead at Accenture.

The systems are integrated with enterprise software, allowing inspection data to be tied to operational workflows, including maintenance reporting and compliance tracking. The approach reflects a broader effort to connect robotics with existing production systems rather than deploying them as standalone solutions.

“Our work in collaboration [with] SAP is a great example of how holistic deployment of humanoid robots... creates a closed loop with transactional systems.” said Prasad Satyavolu, global lead for manufacturing, operations and physical AI at Accenture.

The robots are powered by Accenture’s “Robot Brain” platform and trained using digital twin environments, enabling them to learn tasks through simulation before deployment. The systems are designed to interact with workers using voice, gestures and text, supporting collaborative use cases.

The pilot highlights a growing focus on applying physical AI to inspection and monitoring tasks — areas where variability and real-world conditions have historically limited automation. By combining robotics with real-time data and enterprise integration, companies are exploring how these systems could support safety, quality and operational efficiency in complex environments.

While still in early stages, the initiative reflects broader industry interest in humanoid robotics as a potential tool for extending automation into tasks that require mobility, perception and interaction beyond traditional fixed systems.