ChatGPT is expanding its writing capabilities with a feature designed to make AI-generated drafts feel more consistent with the way individual users already write.

The feature can analyse patterns in a user's work, including vocabulary, tone, sentence construction, preferred phrases and formatting habits, and use those patterns when generating future drafts.

OpenAI's current documentation confirms that ChatGPT Work can work with connected workplace content and create or edit documents using source material and existing templates. OpenAI also supports prompts that ask ChatGPT to review and improve a user's email style.

ChatGPT Learns Writing Patterns

The idea behind the feature is straightforward: instead of asking users to repeatedly describe their preferred writing style, ChatGPT can use examples of their existing work to identify recurring patterns.

These can include word choices, sentence length, punctuation, tone and formatting preferences.

A user who frequently uses particular phrases or structures could therefore receive drafts that more closely resemble their established communication style.

This could be particularly useful for people who regularly produce emails, reports, presentations and other workplace documents.

Gmail And Work Files

The feature is designed around workplace information that users make available to ChatGPT through connected applications and services.

OpenAI's documentation lists workplace integrations such as Google Drive, SharePoint, Box, Notion, Gmail and Outlook Email for various work-related tasks. Users can, for example, ask ChatGPT to analyse their email style in Gmail.

ChatGPT Work can also create and edit documents, spreadsheets, presentations, reports and analyses from instructions and source material. It can be asked to match an existing file, template, layout or tone.

Personalisation Becomes More Powerful

The writing-style capability fits into OpenAI's broader personalisation system.

ChatGPT already allows users to select a personality and adjust characteristics such as warmth, enthusiasm, formatting and emoji usage. These settings work alongside custom instructions and saved memories.

That means style personalisation can operate at several levels. A user can establish a general communication preference through personalisation while also providing specific instructions for a particular document.

For professional writing, OpenAI's “Professional” personality is designed around polished, precise language and conventional workplace writing.

How Style Analysis Works

Large language models can identify recurring linguistic patterns across a collection of writing.

For example, a model can observe whether someone generally prefers short or long sentences, formal or conversational language, active or passive constructions, particular punctuation habits or specific ways of opening and closing messages.

It can then use those patterns as guidance when generating new material.

The important distinction is that the system is not simply copying an earlier document. The goal is to infer stylistic characteristics and apply them to new content.

Useful For Daily Work

The feature could be particularly useful for professionals who produce large volumes of written material.

A communications professional could maintain a consistent voice across emails and reports. A manager could use previous correspondence to generate messages that fit their normal workplace tone. Teams could also use existing documents as references when creating new material.

OpenAI's Work documentation specifically supports creating files from instructions and source material while preserving requirements such as tone, branding, layout, formulas and table structures.

Users Still Control Style

Style matching does not eliminate the importance of user instructions.

OpenAI recommends giving ChatGPT clear and specific instructions and refining prompts when the first result does not meet expectations.

Users can also modify or remove custom instructions. OpenAI says custom instructions can be edited or deleted and apply across chats.

This gives users multiple ways to control how their writing is generated rather than relying solely on automatic style analysis.

A Step Toward Personal AI

The broader significance is that workplace AI is moving from simply generating generic text towards producing content adapted to individual users.

Instead of receiving a standard AI-written email or report, users can increasingly expect drafts that reflect their established vocabulary, tone and formatting preferences.

For professionals who spend significant time writing, that could reduce editing time and make AI-generated content more immediately usable.

At the same time, users should remain aware of what workplace data they connect to AI systems and review organisational privacy and data policies before enabling access to emails, files or other sensitive material.