Research
The Chair of Information Systems and Business Analytics conducts research on the application of machine learning and artificial intelligence (AI) methods in organizations. Our research aims to solve practically relevant problems and to advance the necessary theories and methods. In addition to insights into designing ML-based information systems, we explicitly pursue the development of prototype solutions and bring them into real-world applications to create value for our industry partners. Thereby, we explore the value creation process through AI and advance theories of information systems research.
News
Deniz Oruç Çelik will present initial findings from his study on career success among data scientists at the TDWI Conference in Munich (23–25 June 2026). The contribution, “The Overlooked Profession in the Age of AI? Initial Insights into Data Science Career Success,” examines which observable career factors can explain professional success in the data science occupation. Based on professional networking data, the study shows how established predictors from career research can be used to analyse an emerging IT profession.
Prof. Dr. Konstantin Hopf and Deniz Oruç Çelik will present the talk “Crafting AI: Tensions Between Data Science Craftsmanship and Management Expectations” at data2day 2026 in Cologne. The contribution examines why data science and AI projects in organizations often fail to deliver the expected value. It focuses on the tension between management expectations of technically plannable AI solutions and the craft-based, exploratory nature of practical data science work. Based on more than 100 interviews, the talk presents 15 tactics for more effective management of AI projects.
Die Professur WI2 wird bei der diesjährigen 34. European Conference on Information Systems (ECIS) in Mailand mit zwei Beiträgen vertreten sein.
Ein Beitrag (full paper) zum Vertrauen in Entscheidungen, die von KI-Systemen zusammen mit Menschen getroffen werden. Das Paper „Procedural Justice and the Limits of Explanations in Human-AI Decision-Making“ hat Prof. Dr. Konstantin Hopf zusammen mit Simon Merz (Universität Halle-Wittenberg) und Arisa Shollo (Copenhagen Business School) verfasst. In der experimentellen Studie im ethisch sensiblen Umfeld der Entscheidung über Bewerbungsunterlagen fiel auf, dass Menschen einem Entscheidungssystem weniger vertrauen, wenn Mensch und KI zusammen entscheiden, als wenn nur Menschen oder nur KI-Systeme die Entscheidung treffen. Diese Phänomen beschreiben die Autoren als "hybrid dysphoria", also das Unbehagen bei der hybriden Entscheidungsfindung.
Weiterhin wird Prof. Dr. Konstantin Hopf am Workshop "Teaching with AI in Higher Education: From Insight to Impact“ zusammen mit Leonie Manzke (Universität Erlangen-Nürnberg) mit einem Diskussionspapier teilnehmen. Weitere Autoren in diesem Beitrag sind Prof. Dr. Felix Wortmann und Maximilian May (Universität St. Gallen) und Prof. Verena Tiefenbeck (Friedrich-Alexander-Universität Erlangen-Nürnberg).
Following one semester as acting professor, Prof. Dr. Konstantin Hopf was appointed Chair of Information Systems and Business Analytics, effective February 15, 2026. He thus continues his research and teaching at TU Chemnitz in the areas of explainable machine learning methods and data-driven decision support. Read the full press release.
The article “Validating Explainer Methods: A Functionally Grounded Approach for Numerical Forecasting”, produced in collaboration with the Otto-Friedrich University of Bamberg, will be published shortly in the journal Journal of Forecasting (impact factor 3.2, VHB “B”). The paper focuses on validating explanation methods for numerical forecasting models and shows how the informativeness and reliability of explainable AI approaches for time-series forecasting can be assessed systematically. The publication is Open Access and therefore freely available.
In the February issue of the journal Energy Policy (impact factor 9, VHB “B”), the research article “Do dynamic electricity tariffs change the gains of residential PV-battery systems? A simulation-based evaluation using data from 448 households” will be published. The paper was produced in collaboration with the Otto-Friedrich University of Bamberg and the Julius-Maximilians University of Würzburg. The publication is Open Access and therefore freely available.
The study shows how flexible electricity tariffs can improve the profitability of residential battery storage systems and opens up new perspectives for households and energy policy. The paper has already attracted media attention, including a report in pv magazine on 24 November 2025.