Siemens Revolutionizes Automation with Eigen AI

Siemens Unveils the Eigen Engineering Agent for Industrial Automation

Siemens has recently introduced an innovative artificial intelligence system, the Eigen Engineering Agent. This system is specifically designed to transform the way automation engineering tasks are planned and validated in operational environments.

The Eigen Engineering Agent stands out for its ability to use multi-step reasoning combined with self-correction, enabling autonomous execution of tasks. Directly integrated into engineering platforms, it can manage workflows from the initial design phase to final validation.

An Autonomous Engineering Workflow

This new agent is capable of interpreting project requirements, generating automation code, configuring industrial systems, and refining results until predefined performance goals are met. Among the tasks it can perform are programming programmable logic controllers (PLC), configuring human-machine interfaces (HMI), and setting up industrial devices. It is designed to meet industrial standards for accuracy and reliability.

The Eigen Engineering Agent is connected to the TIA Portal, Siemens' fully integrated automation engineering platform. This connection allows it to access project-specific data, such as structures and relationships between components, enabling it to generate results aligned with existing system configurations, even those that are legacy or undocumented. It can refer to control logic, system hierarchies, and component dependencies, ensuring that results comply with existing engineering standards without requiring manual translation.

The system breaks down engineering problems into steps, processes them sequentially, and evaluates the results against project requirements. It iterates until the results meet the specified criteria before submitting them for an engineer's review.

Addressing Industrial Constraints and Bridging Workforce Gaps

Industry forecasts indicate that there could be a shortage of up to seven million workers in the manufacturing sector by 2030. Some sectors report that about one in five engineering positions remains unfilled. Siemens claims that its system can perform tasks two to five times faster than manual methods while maintaining a high level of accuracy.

During pilot deployments involving over 100 companies across 19 countries, Siemens' system has been applied to standard automation engineering processes. Participating organizations include ANDRITZ Metals, CASMT, and Prism Systems.

Prism Systems utilized the system to generate and import code in Structured Control Language (SCL), significantly reducing the execution time for these tasks. In another instance, CASMT used the system to automate device configuration, code generation, and HMI visualization in production line development, thereby reducing handoffs between specialists in engineering disciplines and shortening delivery times.

The Eigen Engineering Agent is integrated into Siemens' TIA Portal, which has over 600,000 users. It is part of the company's Xcelerator portfolio and is accessible digitally.

Surveys conducted among manufacturing organizations reveal that, while most companies have large volumes of operational data, the quality of this data and its contextualization remain major obstacles.

In addition to general labor shortages, manufacturers also face a lack of workers with the technical skills necessary to operate AI systems in industrial environments.

Initial deployments focus on automation engineering workflows, but the system is structured to expand into other areas of the industrial value chain. Siemens views this development as an integral part of its efforts to integrate AI into industrial operations and software systems.

This initiative follows Siemens' announcement of a 1 billion euro investment in industrial AI. The company states that it employs over 1,500 AI specialists and holds more than 2,000 families of patents related to AI worldwide, thus supporting the ongoing development of AI-based engineering and operational tools.