Rainer Drath (Pforzheim University of Applied Sciences, Tiefenbronner Str. 65, 75175 Pforzheim, Germany)
Presented at
6th AutomationML Conference: 20 Years of AutomationML
22–23 September 2026, Hochschule Pforzheim, Germany
Hosted by Hochschule Pforzheim
Type
Technical presentation — accepted abstract
Open access
© 2026 The Authors. Published under the CC BY-NC-ND 4.0 licence. Peer-review under responsibility of the scientific committee of the AutomationML Conference 2026.
The full contribution will be published on this page after the conference.
Abstract
The AutomationML Editor supports plugins which allow to extend the functionality of the Editor. But development of such plugins requires a rare combination of skills: a) expertise in object oriented AutomationML modelling, b) expertise in AutomationML programming and AML-specific APIs, and c) useful industrial use cases for digital engineering workflows. This combination is rare and requires a steep learning curve in all three dimensions at the same time. Due to a lack of domain-expert contributions, only a few AML Editor plugins are available today, the plugin ecosystem development and innovation speed is very limited.
This contribution aims to significantly reduce the technical development effort for AutomationML Editor plugins, allowing engineers with reduced AML and API know-how to productively produce plugins in short time. This author presents a reproducible, template and AI-driven methodology for AML Editor plugin development. The key idea is to textually explain the AI what is needed instead of programming the code. Prompt based software development is a powerful concept to drastically improve the development performance, estimated by factor 10. The approach lowers technical barriers, empowers domain experts to co-create tooling, and establishes a standardized, reproducible paradigm for AI-augmented industrial software engineering.
To address practical concerns of AI driven software development as software quality and data privacy, the author demonstrates a dual-path workflow leveraging both high-end cloud-based AI coding assistants (e.g. with Claude Code) and locally deployed Large Language Models (with current local LLMs). While the cloud approach enables rapid prototyping and automated code generation with high speed and quality, the local LLM workflow pursues strict data privacy, offline operation, and compliance with sensitive industrial environments, making it suitable for proprietary or regulated engineering contexts. The methodology is structured around prompt engineering, schema-aware code generation. The difference between both approaches is compared and presented in various dimensions as speed and quality.
Furthermore, based on the presented method, this contribution provides a series of new AML Editor plugins for various use-cases. Each plugin is developed, documented, and benchmarked using the proposed AI-assisted pipeline, demonstrating significant reductions in development time and implementation complexity.
Finally, this work positions AI-assisted development as a scalable foundation for next-generation AML tooling and Industry 4.0 data ecosystems. Source code, prompt templates, and usage guidelines are made publicly available later this year.
Keywords: AutomationML; AI-assisted development; Large Language Models; Claude Code; local LLMs; plugin development; industrial software; digital engineering
Speaker
Prof. Dr.-Ing. Rainer Drath is Professor of Mechatronic Systems Engineering at Pforzheim University, Germany. He studied Electrical and Automation Engineering at the Technical University of Ilmenau and earned his doctorate in automation engineering.
Prior to joining academia in 2017, he held various research and leadership positions at ABB Corporate Research, where he worked on systems engineering, virtual engineering, functional safety, and Industry 4.0. Prof. Drath is internationally recognized for his contributions to digital engineering and industrial interoperability. He is one of the principal architects of AutomationML (IEC 62714) and has made significant contributions to the development of CAEX and related international standards. His research focuses on systems engineering, digital twins, Industry 4.0, engineering automation, and model-based engineering. He has authored more than 300 publications and several books and has received multiple international awards for his scientific contributions.
BibTeX
@inproceedings{drath2026aiplugin,
author = {Drath, Rainer},
title = {{AI} based plugin development for the {AutomationML} {Editor}},
booktitle = {Proceedings of the 6th AutomationML Conference: 20 Years of AutomationML},
address = {Pforzheim, Germany},
month = sep,
year = {2026},
publisher = {AutomationML e.V.},
url = {https://www.automationml.org/conferences/conference-highlights/ai-based-plugin-development-for-the-automationml-editor/}
}
How to Cite
R. Drath: “AI based plugin development for the AutomationML Editor”, Proceedings of the 6th AutomationML Conference: 20 Years of AutomationML, Pforzheim, Germany, 22–23 September 2026.
This page presents the accepted abstract of a contribution to the AutomationML Conference 2026. The full text will be added after the conference. If you are an author and would like a correction, please contact office@automationml.org.