KI-basierte Textkreation im Content Marketing: Design und Evaluation eines effektiven Prompts

The challenge in using generative text artificial intelligence (AI) such as ChatGPT is to harness its potential efficiently and optimally in terms of achieving quality goals. In this context, human input to the AI (prompt) is crucial. This article is dedicated to the question of how the AI-based text output of ChatGPT can be specifically controlled by prompt engineering, so that the text quality of the generative AI meets the criteria for successful content marketing texts. The results identify an effective prompt structure for high quality content marketing text copy using ChatGPT. In particular, the Zero-shot Chain-of-Thought, One-shot and Few-shot prompting prove to be successful, as these techniques allow the ChatGPT text output to match the success criteria. In addition, the current weaknesses of AI-generated texts are discussed. This also highlights the limitations of ChatGPT. Through an efficient collaboration between humans and AI these limitations can be overcome and quality goals can be achieved in a joint effort. The theoretically and practically sound results and implications of the research provide guidance to content marketers on the efficient use of ChatGPT.