Claude can now write just like you

AI chatbots are great at generating answers, but they all suffer from the same annoying flaw: that distinct, overly polite, soulless robot voice. Anthropic wants to fix that.

Claude just introduced a new custom response styles feature, allowing the assistant to mimic how you actually write and talk across your conversations.

Training Claude on your own words

Out of the box, Anthropic is rolling out three preset writing styles: Concise, Explanatory, and Formal.

While the presets cover standard office needs, the real highlight is the custom option. Users can now upload chunks of their own writing, whether that means past emails, Slack messages, or full documents.

Claude analyzes your samples and generates a custom persona tailored to your tone. During testing by Tom's Guide, the model analyzed an editor's work and dubbed him a "tech translator" who delivers analytical insights through conversational language.

Best of all, Anthropic says this persona stays consistent across use cases. You won't have to re-explain your preferred writing habits every time you start a new thread.

Why it matters

Every power user has a mental checklist of prompts they paste into every new chat window, usually begging the AI to stop using fluffy adjectives and robotic corporate speak. Building writing personas directly into the UI eliminates that friction entirely.

We think this is a smart move by Anthropic. As raw intelligence among frontier models continues to plateau, the battle is shifting toward personalization and workflow retention. OpenAI has relied on broad custom instructions and general memory, but Claude's laser focus on matching voice and tone makes the assistant feel noticeably more human.

That said, matching someone's voice from a static writing sample isn't foolproof. A single tone might work for team updates, but fail miserably for delicate client emails. Still, giving Claude a reliable memory of who you are makes it significantly more practical for daily work.