On October 20, 2026, Philipp Reinhard successfully defended his dissertation titled“From Generative to Agentic AI: A Processual, Interactional, and Technological Perspective on Augmentation.”

During his dissertation defense, he presented key findings from his research on the question of what patterns of augmentation emerge when generative AI (GenAI) is integrated into service routines, and how these patterns influence the delivery of services. The focus was on a processual perspective at the micro level, which he used to examine how generative AI reconfigures routines both in the “frontstage”—that is, in direct customer contact—and in the “backstage”—that is, in the supporting processes behind the scenes.

His findings show that seven augmentation patterns (MAPs) can be distinguished, which refine the continuum between substitution and complementation at the level of individual routines. Furthermore, the study identifies service permeation mechanisms that illustrate and explain how augmentation affects service delivery. It also describes documentation and the targeted provision of information to AI (AI feeding) as a new category of routines in service delivery.

With his dissertation, Philipp makes both theoretical and practical contributions to research at the intersection of business informatics, service research, and artificial intelligence. Going beyond the process-oriented perspective, he expands augmentation research at the interactional level in his dissertation by identifying the conditions under which interaction between humans and generative AI succeeds—whether during AI training, the use of copilots, or with regard to explainability. At the technological level, he develops a conceptualization of agent-based AI and collaboration in multi-agent systems, thereby expanding the understanding of delegation beyond human-AI dyads to constellations of multiple agents.

His research findings also yield concrete recommendations for practice. Organizations should begin with use cases such as summarization, copiloting, and coaching, and view generative AI as an active participant in service delivery, not merely as an automation tool. In addition, they should regularly assess where and to what extent generative AI permeates both front-stage and back-stage operations in order to identify gaps in integration at an early stage. At the same time, service employees must retain responsibility: establishing safeguards against excessive reliance on AI, as well as building competencies in working with AI and the judgment to know when to override AI recommendations, are central to this. Furthermore, the study offers guidelines on how to involve employees in AI training, select “copilots” appropriate to the difficulty of a task, and appropriately calibrate trust in AI through explanations. For the design of agent-based AI, it also provides a governance blueprint that enables such systems to be evaluated, designed, and orchestrated, and which explicitly addresses the balance between autonomy and oversight.

The Department of Business Informatics warmly congratulates Dr. Philipp Reinhard on successfully defending his dissertation and wishes him all the best and continued success in his future endeavors.