I’d like to learn how to use Spiral more effectively.

I’m a Korean user currently working in Korea as a content planner and editor. I’d like to use Spiral to generate first drafts that accurately reflect more than 90% of my writing style. However, after several tests, it seems that the model doesn’t fully understand the “style” sample I provided for reference. The drafts often contain awkward literal translations from English and unnatural, machine-like sentence transitions.

Is this happening because I’m writing in Korean? If Spiral can outperform Claude in this area, I’m very interested in adopting it as my main tool. Could you please advise me on how I, as a Korean user, can make the most out of this model?