Sounds interesting? Here's the specific research question:
LLMs as Intelligent User Interfaces for Complex Optimization and Analysis Software
- Traditional chatbot interfaces allow for free-form text input but do not adequately reflect the actual usage contexts of complex specialized software. It remains unclear which forms of interaction are better suited for demanding analytical tasks.
- In particular, the question arises as to whether LLM-based interfaces should not only respond reactively to user input, but also proactively provide hints, ask follow-up questions, or offer recommendations for action.
- The study will examine how different interaction paradigms—such as chat, form-based navigation, generated UI elements, or proactive assistance—affect usability, trust, speed, and the quality of results.
Here is how we envision this methodology:
- Development and comparison of several LLM-based interface concepts for a complex software application in the OPTANO context.
- Implementation of a prototype with at least two interaction modes, e.g., traditional chat versus proactive, context-aware assistance based on usage or system data.
- Evaluation through user testing and qualitative interviews, as well as quantitative metrics such as completion time, error rate, perceived usefulness, and sense of control.
You'll need the following knowledge:
- Programming experience, ideally in C#
- Basic knowledge of human-computer interaction, UX, or empirical evaluation
- Basic Understanding of LLMs, Prompting, and Agent-Based Systems