Eshima, Imai, and Sasaki (2023) propose a significant improvement in the measurement performance of topic models by incorporating a limited set of keywords. Their research introduces an advanced topic model that effectively aggregates these keywords. However, I argue that the development of this new model is unnecessary. Instead, scholars can conveniently utilize keywords to cluster documents based on a large language model. My research demonstrates that the approach using a large language model is not only considerably faster but also more accurate compared to the keyword-assisted topic models. Therefore, I urge scholars to embrace this new method as a preferable alternative to continuously modifying outdated approaches.