Agentic AI for Research Workflows: A Human-Centered Approach
Requirements
- Required: Successful participation in the course "Human-Computer Interaction"
- Desirable: Successful participation in the seminar on "Interactive Intelligent Systems" and the lecture on "Wissenschaftliches Arbeiten in der Informatik"
- Advanced Programming knowledge (Python)
- An understanding for Machine Learning and AI concepts
- Interested in conducting user studies
Contents
1) Content
How can agentic AI systems be designed to meaningfully support researchers in their daily workflows?
Agentic AI systems are characterized by autonomy, goal-directed behavior, and the ability to act proactively in complex environments. In research contexts, such systems have the potential to support information management, collaboration, and sensemaking. However, current approaches often fail to adequately reflect the dynamic and situated nature of real-world research practices.
This thesis investigates how agentic AI systems can be designed to provide meaningful and context-sensitive support for researchers. The focus lies on understanding everyday practices, identifying opportunities for proactive assistance, and translating these insights into system design.
The aim is to bridge the gap between recent advances in agentic AI and the practical requirements of academic work through the design and evaluation of a prototype, grounded in a human-centered design approach.
2) Research Approach
The thesis follows a human-centered process combining qualitative research, system development, and evaluation.
First, semi-structured interviews with researchers are conducted to understand workflows, challenges, and expectations. The collected data is analyzed using qualitative methods to derive structured system and user requirements.
Based on these requirements, an agentic AI prototype is designed and implemented.
Finally, the system is evaluated in a thinking-aloud study. User interactions are analyzed to identify usability issues, assess workflow integration, and evaluate the overall utility of the system.
References
1) Hornbæk, Kasper, Per Ola Kristensson, and Antti Oulasvirta, Introduction to Human-Computer Interaction (Oxford, 2025; online edn, Oxford Academic, 21 Aug. 2025), https://doi.org/10.1093/oso/9780192864543.001.0001, accessed 25 Mar. 2026.
2) Joon Sung Park, Joseph O'Brien, Carrie Jun Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein. 2023. Generative Agents: Interactive Simulacra of Human Behavior. In Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology (UIST '23). Association for Computing Machinery, New York, NY, USA, Article 2, 1–22. https://doi.org/10.1145/3586183.3606763
3) Gomez C, Cho SM, Ke S, Huang C-M and Unberath M (2025) Human-AI collaboration is not very collaborative yet: a taxonomy of interaction patterns in AI-assisted decision making from a systematic review. Front. Comput. Sci. 6:1521066. doi: 10.3389/fcomp.2024.1521066