Siemens Expands Generative AI-Powered Industrial Copilot

The Industrial Copilot accelerates engineering and shopfloor operations.

The Industrial Copilot helps thyssenkrupp engineers to create a machine visualization in WinCC Unified.
The Industrial Copilot helps thyssenkrupp engineers to create a machine visualization in WinCC Unified.
Siemens

Siemens announced new functionalities for the Industrial Copilot, a generative AI-powered assistant for engineering in an industrial environment, and added thyssenkrupp Automation Engineering as global customer.

The Siemens Industrial Copilot for Engineering currently writes code for automation engineering and future capabilities include multimodality and agent concepts. To deliver full data sovereignty, the Siemens Industrial Copilot for Operations is planned to be offered as an on-premises hardware-software bundle. 

thyssenkrupp Automation Engineering and Siemens Electronics Factory to roll out the Siemens Industrial Copilot

thyssenkrupp Automation Engineering, a special machine and plant builder, integrated the Copilot for Engineering in a battery machine used for battery quality inspections on electric cars. The industrial company plans to use the genAI-powered assistant at scale – engineering the machines at thyssenkrupp’s global locations from 2025 onwards.

The Industrial Copilot assists thyssenkrupp engineers in creating TIA Portal projects. It helps them develop structured control language (SCL) code faster for programmable logic controllers (PLCs), intelligently integrates the code into the TIA Portal and generates a machine visualization in WinCC Unified. This allows engineering teams to reduce repetitive and monotonous tasks like automating data management and sensor configuration. 

The Siemens Electronics Factory in Erlangen, Germany, implemented the Copilot for Operations across its soldering machines. The Industrial Copilot helps Siemens operators and maintenance engineers to understand a machine’s error codes by translating its messages into natural language.

It suggests solutions based on the machine’s details and history by combing through different documents, manuals and spare part lists. Machine downtime can be significantly reduced, production bottlenecks can be resolved faster and shift handovers will work more efficiently.

Multimodality, agent concepts and on-premises approach to supercharge the Siemens Industrial Copilot

The Industrial Copilot for Operations allows shopfloor workers to directly interact with machines and helps them with maintenance tasks, error handling and performance optimization. Additionally, the Industrial Copilot will have multimodal capabilities to analyze and interpret images and drive even more productivity with agent-based automation for a variety of tasks.

To address data security for customers and make sure that data doesn’t leave the shopfloor, the Industrial Copilot for Operations is planned to be offered as an on-premises hardware-software bundle with the Simatic Industrial PC (IPC 1047E). The software stack running on IPCs is powered by NVIDIA NIM microservices, part of the NVIDIA AI Enterprise software platform, which lets automation and maintenance engineers ask real-time queries about operational and document data to facilitate rapid decision-making and reduce machine downtime.

This configuration doesn’t require an Internet connection and stores data on local hardware devices. It helps ensure data security by processing all data right on the shopfloor and keeping customers’ data stored and available when and where it’s needed. 

The Industrial Copilot for Engineering will support multimodal input, for instance, by detecting and converting manual changes in the ECAD document that’s used for electrical planning. These changes are automatically highlighted, annotated and eventually implemented in the TIA Portal project. 

Highly complex automation projects will be partially automated using agent concepts. Agent concepts go beyond simple question-and-answer interactions, automating processes by breaking down large, complex tasks into subtasks.

All relevant information is then collected from a number of sources, including ECAD information, in order to understand the user goal. Agents can also be connected to external systems and sources, which creates a closed loop with different tools linked together.

Next, the agents create a plan on how to achieve goals and execute the required actions independently. These range from sending messages and accessing external systems to updating data sets. Engineers can also use agents to control and direct all production processes  while maintaining full transparency, having an overview of the data and knowing which steps should be taken next. 

The Engineering Copilot TIA Essential has been available on the Siemens Xcelerator marketplace since July 2024. While Siemens provides the automation elements of the Industrial Copilot, the natural language processing is carried out by one of the most powerful GPT models using the Azure OpenAI Service of the Microsoft Cloud.

More in IoT