Economists constantly process large amounts of information from newsletters, blogs, websites, official announcements, central bank publications, speeches, and market reports. A key challenge is to identify which sources are worth reading in detail, which information is genuinely new, and how relevant insights can be transformed into forecasts, reports, and publications.

This master’s thesis addresses this challenge by designing and evaluating a human-AI tool that supports economists in monitoring, selecting, and pre-assessing relevant economic information. The goal is not to replace expert judgment, but to make the information selection and analysis process more efficient and convenient.

The thesis will be conducted in collaboration with Deka, which is part of the Sparkassen-Finanzgruppe. The application context is the work of economists who create economic analyses, forecasts, market assessments, and publications for different internal and external audiences. The thesis is embedded in the Flow Factory, a joint research lab between the University of Münster and the Sparkassen-Finanzgruppe. Within the Flow Factory, the thesis contributes to ongoing practice-oriented Information Systems research on AI-supported knowledge work and human-AI collaboration.

The thesis will follow a Design Science Research approach. You will analyze current information routines, work closely with practitioners, and develop a working prototype. The tool should allow users to define relevant sources, detect new content, generate German-language summaries, assess relevance, and support decisions about which documents should be read in full.

Beyond developing a proof-of-concept, the thesis should also demonstrate proof-of-value by evaluating whether the tool creates practical value for economists in their daily work, for example by making source selection more convenient, improving prioritization, or enabling users to process a larger information funnel more efficiently.

Possible guiding research question:

• How can a human-AI tool be designed to support economists in monitoring, selecting, and pre-assessing relevant economic information?
• Expected tasks include conducting interviews with economists, analyzing typical input sources and output examples, designing a human-AI workflow, implementing a working prototype, evaluating its practical value with users, and deriving design implications for AI-supported expert knowledge work.
• The thesis is suitable for students interested in generative AI, human-AI collaboration, knowledge work, and design-oriented research. Programming skills for prototype development are required.