Indirect Value Creation Potentials of AI Application in Companies

This study explores questions concerning AI value creation potentials in companies and, in particular, focuses on indirect value creation potentials. In order to adequately address this complex topic, more than 90 AI use cases were collected and evaluated in over 40 qualitative interviews with experts from 34 companies. For an in-depth survey of the indirect value creation potentials, open questions were posed in the interviews. These questions sought to find out which goals were being pursued through the application of AI. The evaluation was carried out using a qualitative content analysis based on Mayring ([1] 2016). Subsequently, all statements on indirect value creation potentials were assigned to inductively formed categories. This showed that the recorded and systematised goals of AI applications contribute indirectly to the value creation of the companies and go beyond the direct value creation factors of reducing costs, increasing sales and increasing company value ([3] Seitz/Willbold/Haiber, 2022). In the evaluation, general statements were made about indirect value creation potentials across industries and companies. A total of 14 indirect value creation categories emerged, for example optimising decision-making, but also increasing the productivity of employees. In a more detailed analysis, it was determined whether and to what extent there were differences, based on the industry and the company's field of activity. The goals pursued with the application of AI in companies differ depending on the application scenario. While companies with a high level of customer contact tend to see potential in optimising the customer journey, manufacturing companies, for example, are more concerned with process optimisation. In a next step in the research, it would be useful to develop a more complex evaluation system for value creation potential in order to give companies an indication of rewarding AI projects. The indirect value creation potentials identified in the study can provide an initial starting point for this.