A previous thesis at the chair for Software Engineering developed a prototype of a Text-To-Process Model Generator for the subject-oriented process modeling language PASS.  The main focus of that work was to create an Architecture and a running prototype overall.

This goal has been achieved and the software is available at: https://github.com/I2PM/Text-to-PASS-Model-Generator

The goal of this Thesis is to evaluate the software by applying increasingly complex processual descriptions to answer the following questions and goals:

• What are the limits of the Software (size of possible Input/complexity of a described process, imprecise textual descriptions, different languages, etc.)?
• If Limits are found, are there ways to overcome those (Better Prompts? using different LLMs? Better User Instructions for Users?)
• What is the usage process, and can it be improved? Interactions?