As a valuable tool for tourism, sentiment analysis requires in-depth knowledge of linguistics and research context, as well as substantial time and monetary investments. The existing tools aiming at sentiment analysis automation generally lack validity compared to the manual analysis. However, contemporary technologies create new opportunities for Natural Language Processing and, therefore, sentiment analysis. This paper evaluates the effectiveness of GenAI for sentiment analysis and benchmarks it against two alternative coding approaches—manual coding and NVivo. The paper uses 301 online customer reviews of hotels implementing service robots in their front-end operations. The results show that ChatGPT systematically outperforms NVivo but falls short compared to manual coding. Both ChatGPT and NVivo correctly identify positive and negative emotions in most reviews. However, both fail to distinguish the degree of polarity and avoid extreme sentiments (very positive/negative). Further investigation and improvement of sentiment analysis tools are required.