Portrait of Friedrich Stallinger

Process Management

by Friedrich Stallinger

University of Applied Sciences BFI Vienna UAS BFI Vienna

Master · 3 ECTS · 1st semester · 67 students

AI was integrated across four connected assignments over roughly three months. Students first used a customized chatbot as a tutor for process lifecycle models, then worked with prescribed GenAI tools on process landscapes, BPMN-based process modelling, and investigations of technology trends in process management. The primary goal was to combine authentic, practice-oriented process work with critical reflection on AI use and AI output quality. What set the design apart was bounded comparability: two lecturer-written company cases, prescribed tools, and cross-group peer review made differences in AI use discussable.

Learning outcomes

  1. Describe selected process lifecycle models and explain their principles, structure, and context of application
  2. Acquire, model, and document processes using a standard notation
  3. Describe current trends in process management and interpret their effects on process-oriented organisations
  4. Analyse and improve existing processes and work out key indicators for process measurement and controlling
  5. Explain and distinguish process reference, process maturity, and process capability models
  6. Relate process management to process-oriented quality management and integrated management systems
  7. Develop a process landscape in line with corporate strategy in simplified business contexts

Assessment

Students received points for successful AI-related tasks and submitted written reports and reflections on all AI-related assignments via Moodle. Assessment was mainly based on qualitative criteria highlighting students' own work: depth of investigation and insight, critical evaluation of AI outputs and AI tool usage, and documentation of AI usage. The lecturer gave feedback and up to 5 points for each submission, according to the respective assessment criteria. These points contributed to the final grade but did not replace the final exam, which was required to successfully complete the course.

Evaluation

Impact was evaluated through the pilot logbook, in-class discussions after assignments, feedback gained from peer-review activities, written reflections and analyses, and a separate informal evaluation conducted by the lecturer about the CustomGPT tutor. In addition, a comparison of overall performance with previous years was conducted to evaluate the impact of AI on the learning outcomes.

  • KPIs tracked: Yes — General KPIs: performance (grades) compared to previous years; level of engagement during in-class assignments and discussions; and the kinds of questions students ask about the topic, assignments, or AI tools and usage. KPIs for written reflections varied by task — e.g. clarity and comprehensiveness of the lifecycle comparison and quality of AI tool evaluation for the self-study task; depth of investigation, presentation quality, and critical reflection for the trends task; alignment and coherence of the landscape and relevance of peer-review feedback for the landscape-development task; and correctness, clarity, and critical discussion of limitations for BPMN modelling. AI usage documentation was tracked across all tasks.
  • Formal institutional evaluation: No

Risk management

Risk management worked mainly through didactic design rather than formal compliance procedures. Introducing the customized chatbot explicitly as a tutor instead of a ghostwriter, and explaining the purpose of the task, mitigated the risk of students handing in AI-generated texts instead of their own comparison of process landscapes. To ensure comparability, specific company cases and AI tools were prescribed, and students had to document their AI usage; the peer review activity also surfaced differences in results and respective prompts. Unforeseen risks emerged during the semester nonetheless: students sometimes used tools that were not prescribed or used them in unintended ways, uploaded course materials and literature without reflecting on legal/property-right issues, and partly relied on AI in ways that could undermine hands-on competencies. These were mitigated by addressing the issues in class and explaining the purpose of tasks to students.

Challenges

Scale and workload — two large groups meant many AI tasks had to happen outside of class, reducing lecturer oversight and collective reflection time; preparation was work-heavy because customized chatbots, company cases, licences, new AI tool developments, and copyright/material issues all had to be handled
Managed through prescribed cases and tools, structured peer review, guiding questions, and written feedback
Students sometimes used AI beyond the intended tutor function, uploaded teaching materials or literature without reflecting on legal implications, and faced technical problems around BPMN data/exchange formats
Addressed through explicit in-class discussion of limitations and by reinforcing the purpose and boundaries of each AI task

Scalability

The experience is transferable to other applied management or operations courses that use case work, comparative outputs, and reflection on AI-supported problem-solving. Its strongest transferable elements are the bounded design (prescribed cases/tools), the tutor-chatbot for self-study, and the cross-group peer-review format. To scale it well, the institution would need smaller class sizes, fewer AI-supported assignments and respective licences, more time for in-class debrief, legal and ethical guidance on uploads/materials prior to the course, and allocated time for developing and updating custom chatbots and class materials. Without this support, the design becomes too uncontrolled and too labour-intensive.