OpenAI is testing a ChatGPT Writing Style feature that reads your existing messages, documents and emails from connected workplace apps, then uses those patterns to write new content in your voice. The onboarding screen puts it plainly: ‘ChatGPT will write in your voice by referencing examples from your connected apps.’
The feature is currently available to a limited group of users. Apps listed in the integration include Slack under messaging, Google Drive and Notion for documents, and Gmail for email. TechDogs also notes SharePoint as a supported source, extending the reach into Microsoft’s ecosystem. The idea is that ChatGPT builds its understanding of how you communicate across different contexts, the clipped register of a Slack message versus the more considered tone of a long-form document, and carries those signals into whatever it drafts next.
What the ChatGPT Writing Style Feature Actually Picks Up
This goes beyond broad tone matching. According to Superpower Daily, the feature can detect preferred phrases, capitalisation habits, formatting choices and sign-offs, then apply those patterns when generating content. If you habitually skip the Oxford comma, write in sentence case, and sign off with a breezy ‘Cheers’ rather than a formal ‘Kind regards’, ChatGPT should, in theory, replicate all of that without you ever having to explain it.
The patterns learned apply to content generated in ChatGPT Work on both the web and mobile, meaning the consistency follows you across devices. Setup happens on the web, through Settings, then Personalization, then Writing Style. The feature is available to customers on all paid ChatGPT plans, not just enterprise tiers.
OpenAI has confirmed it is testing Writing Style, but has not said when it will roll out more widely.
How This Compares to Claude’s Styles Approach
The closest point of comparison in the market is Anthropic’s Styles feature, which lets users upload or paste writing samples and then generates a custom style based on those communication patterns. OpenAI‘s approach takes a different route: rather than asking users to curate and submit examples manually, it pulls from writing that already exists inside connected services.
That distinction matters in practice. Manual uploads put the burden of curation on the user, you have to decide which emails represent your style, copy them, and paste them in. Referencing live connected apps means ChatGPT can draw on a much larger and more varied corpus, including how your register shifts between a quick Slack reply and a formal client proposal. The range of real examples is wider, and the setup friction is lower.
Whether that broader access also raises more eyebrows about data handling is a fair question, and one OpenAI has not yet addressed in detail publicly. Granting an AI assistant read access to your full Gmail history and Google Drive is a different proposition from pasting in three writing samples yourself.
The Bigger Picture for AI Writing Assistants
The ChatGPT Writing Style feature fits into a broader push by AI assistant developers to reduce the repetitive setup cost that currently frustrates power users. At present, anyone who wants ChatGPT to write in a particular voice typically has to paste in examples, describe their preferences, or set up a custom system prompt, and repeat some version of that in every new conversation. A persistent, app-connected style profile sidesteps that entirely.
For workplace users especially, the appeal is obvious. Drafting a routine email or summarising meeting notes in your own voice rather than in generic AI prose is a real, everyday use case. Whether the feature delivers reliably across the messiness of real inboxes and document histories is what the current limited rollout is presumably meant to establish.
OpenAI has not announced a general availability date, so users on paid plans who want early access should check Settings under Personalization to see whether the Writing Style option has appeared for their account yet.
