Leverage Data Streams For Better Operational Decision Making

Leverage Data Streams for Better Operational Decision-Making

Smart sustainable mobility ecosystems promise to address society’s expectation of environmentally friendly on-demand mobility. While the technology stack to build such ecosystems is just around the corner in the form of connected, automated, and electric vehicles, strategies to deploy and operate such fleets in a coordinated manner must still be advanced. Most of such optimization challenges highly depend on the nature of customer demand, vehicle supply, and environmental influences. Hence, this dissertation investigates how available data streams from mobility ecosystems can be leveraged in Information Systems to solve related decision problems. The overarching goal of this work is to generate design knowledge to improve vehicle availability, provider profitability, and environmental sustainability for such ecosystems. Applying quantitative methods to real-world data from shared vehicle systems generates insights into the nature of demand and supply. Combining it with an analysis of empirical research on vehicle relocation algorithms builds the foundation for two artifact designs. The first artifact enables the development and simulation-based evaluation of operation modes for vehicle fleets. The second artifact enables artificial intelligence-based decision support for the vehicle rebalancing problem. The insights are finally incorporated and generalized to a nascent design theory on data-enabled operational decision-making in the context of smart sustainable mobility environments. The findings have multifaceted implications for researchers concerned with data-enabled value creation in Green IS, shared economy and smart mobility, and business analytics and data science. Furthermore, guidance for fleet providers to improve system attractiveness and for society to experience the potential amount of vehicle access without personal ownership is provided.

Leverage Data Streams for Better Operational Decision-Making

Aligning Artificial Intelligence with Organizational Contexts (Band 123) An Investigation of Natural Language Processing in Clinical Healthcare

Künstliche Intelligenz (KI) bietet enormes Potenzial, um Herausforderungen wie Fachkräftemangel und eine wachsende Zahl älterer Patientinnen und Patienten zu bewältigen. Gleichzeitig jedoch stellt die soziotechnische Einbettung von KI in Organisationen eine komplexe Aufgabe dar, deren Misslingen zu Wissensverlust oder Unzufriedenheit im Team führen kann. In dieser Dissertation werden die entscheidenden Komponenten dieses KI-Alignments systematisch untersucht. Vier Studien – davon zwei auf allgemeiner Ebene und zwei mit Fokus auf Natural Language Processing (NLP) im klinischen Bereich – beleuchten, wie KI und organisatorische Anforderungen bestmöglich aufeinander abgestimmt werden können. Dabei entsteht ein Rahmenwerk, das sowohl Forschenden als auch Praktikerinnen und Praktikern hilft, KI-Lösungen gezielt zu planen, zu integrieren und kontinuierlich zu verbessern. Zugleich eröffnet die Arbeit neue Perspektiven für Forschung und Anwendung, vor allem durch die Betrachtung von KI in spezifischen Kontexten und durch die gezielte Identifikation bisher unbeachteter Aspekte. Artificial Intelligence (AI) offers enormous potential for tackling challenges such as staffing shortages and a growing elderly population with increasing healthcare needs. However, effectively embedding AI in organizations is a complex sociotechnical endeavor; if misaligned, it can lead to knowledge loss or employee dissatisfaction. This dissertation systematically examines the key components of AI alignment. Four studies—two at a general level and two focused on Natural Language Processing (NLP) in clinical settings—explore how AI can best be harmonized with organizational requirements. The resulting framework supports researchers and practitioners alike in planning, integrating, and continuously refining AI solutions. Moreover, this work opens new horizons in research and practice by examining AI in specific contexts and identifying aspects that have received little attention thus far.