Anwendungspotentiale von Large Language Models im Innovationsprozess

  • Hugo Maximilian Wagner

Student thesis: Master's Thesis

Abstract

In an increasingly dynamic and complex business world, companies are faced with the challenge of continuously innovating and improving their products, services and processes in order to remain competitive. Innovation processes therefore have to be run through more frequently and product life cycles are becoming shorter. Against this background, this master's thesis analyses the potential of Large Language Models (LLMs) in innovation processes. The aim of this thesis is to analyse and evaluate the possible applications of this advanced artificial intelligence in various phases of the innovation process by going through an innovation process with an LLM in the empirical part. At first, the theoretical foundations of artificial intelligence and LLMs as well as innovation processes are explained. On the basis of these theoretical foundations, an innovation process model is selected for the empirical part of the work, whereby the selection is based on defined criteria. The specific methods used in the innovation process of the thesis are selected according to the chosen innovation process model. In order to utilise the most powerful LLM currently available in the research of this thesis, a model is chosen that has been rated as leading by independent evaluations. To determine the current state of research and the practical application of LLMs in innovation processes, the status quo in scientific research and the economy is analysed in a chapter. This is followed by an empirical investigation in which an LLM is used in a case study in an innovation process in order to explore its performance and benefits. The results of the case study show that LLMs, such as the LLM GPT-4o used, are able to significantly support the innovation process by making it possible to organise the innovation process much more efficiently through fast responses and thereby generate high-quality results. In addition, the innovation process shows that LLMs promote creativity by combining existing knowledge in new ways and thereby creating innovation. Throughout the entire in-innovation process, the LLM offers more than just support and runs through innovation process phases completely independently. At the same time, however, challenges also become clear, such as the dependence on the quality of the prompts and the need for continuous further development of the models. The investigations in the empirical part of the thesis lay the foundation for further research and show that LLMs have the potential to support innovation processes. Existing studies focus too much on individual methods, but the innovation process should be considered holistically in future studies.
Date of Award2024
Original languageGerman (Austria)
SupervisorPatrick Brandtner (Supervisor)

Cite this

'