I’ve found Gemini and Codex can do pretty well by just telling them to follow my writing style, be brief, be less robotic, be professional, be semi-professional, use less technical jargon, no em-dashes, etc. depending on the task. But, I only use that for short summaries of things I am deeply familiar with to send a slack message about or similar. I also edit before sending. It takes maybe a couple of prompts to get 90% of the way there.

Claude on the other hand is atrocious at sounding human, it’s like it’s deliberately following LLM -speak stereotypes no matter how you tell it to write. I finally told it to stop using “Claudeisms,” which gave better though not great results.