Exploring the Use of Agentic AI Tools in Automation Engineering: The Case of Siemens Industrial Copilot
Melnichikhina, Ekaterina (2025)
2025
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2025120833411
https://urn.fi/URN:NBN:fi:amk-2025120833411
Tiivistelmä
This thesis explores the integration of Siemens Industrial Copilot into the design processes of LSK Technology Oy and examines how engineers perceive its usability, efficiency, and trustworthiness. The study applies a constructive research approach and includes semi-structured interviews with five design engineers before and after a short pilot conducted in TIA Portal with Step 7 Professional and WinCC Unified. The aim is not to provide recommendations for improving Copilot itself, but to evaluate its capabilities, limitations, and its potential role in LSK’s engineering workflow.
The findings show that Siemens Industrial Copilot Version 1 provides value in SCL code generation, block and datatype creation, and documentation search, but is not yet reliable for HMI development, hardware configuration, or projects based on Ladder/FBD languages. Participants’ trust in the tool's behaviour largely depended on its accuracy and transparency and remained relatively stable before and after the trial. Engineers’ perceptions were mixed: some participants were disappointed by the tool’s limitations, while others accepted its early-stage imperfections and remained satisfied with its useful features.
The study contributes to understanding industrial copilots within current AI developments and applies Human-AI interaction, usability, and trust perspectives in an industrial context. Recommendations support the future adoption of AI tools at LSK Technology Oy and may also be applied in other organisations and projects working in a similar field.
The findings show that Siemens Industrial Copilot Version 1 provides value in SCL code generation, block and datatype creation, and documentation search, but is not yet reliable for HMI development, hardware configuration, or projects based on Ladder/FBD languages. Participants’ trust in the tool's behaviour largely depended on its accuracy and transparency and remained relatively stable before and after the trial. Engineers’ perceptions were mixed: some participants were disappointed by the tool’s limitations, while others accepted its early-stage imperfections and remained satisfied with its useful features.
The study contributes to understanding industrial copilots within current AI developments and applies Human-AI interaction, usability, and trust perspectives in an industrial context. Recommendations support the future adoption of AI tools at LSK Technology Oy and may also be applied in other organisations and projects working in a similar field.