In a joint project, the University of Ulm, in collaboration with Liebherr-Digital Development Center GmbH, has investigated how the use of GenAI affects productivity and outcomes in knowledge work. GenAI systems are increasingly being used in everyday work, for example to research information, structure content or develop new ideas. Many companies expect this to lead to significant productivity gains. At the same time, there is currently only limited knowledge of how GenAI actually affects the efficiency, quality and diversity of the results of knowledge work. Against this backdrop, a user study was conducted in collaboration with Liebherr, in which employees carried out typical knowledge work tasks – including information research, summarising content and developing new ideas. Some of the participants were supported by a GenAI application, whilst a control group carried out the same tasks without such support. The results show that GenAI can significantly speed up the completion of many tasks. At the same time, the impact on the quality of the results depends on the nature of the task. Employees with lower baseline performance benefit particularly clearly, whilst the performance gap between so-called top and bottom performers is partially reduced by GenAI support. Furthermore, the results suggest that GenAI-supported knowledge work can lead to more similar outcomes. A conscious and reflective approach to GenAI – for example, through targeted prompts and critical further processing of the generated content – can help to reduce such effects.

Cooperation partner: Liebherr-Digital Development Center GmbH

Project period: 2024 – July 2025