New Delhi: OpenAI has introduced a new Writing Styles feature for ChatGPT Work that allows the AI assistant to learn how a user writes and reproduce elements of their personal writing style in future drafts. The feature can analyse writing from connected services such as Gmail, Google Drive, Slack and SharePoint, allowing ChatGPT Work to understand a user’s preferred words, tone, punctuation and other writing habits.
The move could make AI-generated emails and other professional communications sound more personalised. Instead of relying only on generic instructions such as “make this more professional”, users can allow ChatGPT Work to study their existing writing and use those patterns when generating new content.
However, the feature also raises questions about privacy and how much personal communication users are willing to make accessible to an AI system.
ChatGPT Work can learn from your writing
OpenAI announced the Writing Styles feature through the official ChatGPT account on X. The company says ChatGPT Work can identify characteristics that make a user’s writing distinctive, including favourite phrases, specific sign-offs and even unusual capitalisation habits.
This means the system can go beyond simply following a broad instruction such as “write in a friendly tone”. It can attempt to reproduce the patterns that make a particular person’s communication recognisable.
For example, someone who regularly uses a particular greeting, prefers short sentences, uses certain expressions or signs off emails in a specific way could see those characteristics reflected in AI-generated drafts.
The objective is to make AI-written content feel less generic and closer to the user’s established communication style.
Users need to connect their everyday apps
Writing Styles is not based solely on a short writing sample entered into a ChatGPT conversation.
According to India Today, users need to connect applications they regularly use, including Gmail, Google Drive, Slack and SharePoint, to ChatGPT Work. Once those services are connected, ChatGPT Work can access writing from emails and messages to identify patterns in how the user communicates.
This could make the feature considerably more useful for professionals who write large numbers of emails, internal messages, reports or other business documents.
At the same time, connecting communication and productivity services gives the AI access to a broader set of information than a conventional chatbot prompt would contain. Users therefore need to understand what services they are connecting and what information those services may contain.
How does the AI learn your writing style?
OpenAI has not publicly provided detailed technical information about exactly how Writing Styles analyses a user’s communications.
The feature is expected to rely on language-processing techniques to identify recurring patterns in a user’s writing. These can include vocabulary, sentence construction, punctuation, tone and formatting habits.
The system can then use those patterns when producing subsequent drafts.
Importantly, the feature is about style imitation, rather than simply copying previous emails. The purpose is to allow ChatGPT to generate new text while reflecting characteristics associated with the user’s existing writing.
That could be particularly useful for people who want AI assistance without ending up with messages that have an obviously generic AI-generated tone.
Writing Styles is currently limited to ChatGPT Work
One important limitation is that Writing Styles is not being presented as a standard feature for every ChatGPT user.
India Today reports that the capability is currently limited to ChatGPT Work, OpenAI’s agentic AI tool. OpenAI has not announced whether the same feature will be made available to the standard ChatGPT chatbot.
The feature is available on the web for paid plans that have access to ChatGPT Work. Users can set it up through the web interface or through Settings under the Personalisation options.
This makes the launch more relevant to professional users and organisations using ChatGPT for workplace productivity.
OpenAI is not the first to offer personalised writing styles
The concept is not entirely new in the AI industry.
Anthropic previously introduced a Styles feature for Claude, allowing users to upload or paste writing samples and create a custom style based on their communication patterns.
The growing competition between AI companies is pushing assistants beyond simple question-and-answer interactions. Personalisation is becoming an increasingly important part of AI products, with companies attempting to make their systems understand not just what users ask for, but also how they prefer information to be presented.
For workplace users, writing-style personalisation could reduce the amount of editing required after an AI generates an email or document.
Privacy becomes a key consideration
The convenience of the feature comes with an important trade-off.
To allow ChatGPT Work to understand an individual’s writing style, users need to give it access to connected services containing their previous communications. Depending on the account, these services could contain professional correspondence, internal messages, documents and other sensitive business information.
That means users and organisations should carefully review permissions before connecting workplace accounts.
Companies may also need to consider their own data-governance policies before employees connect corporate email, messaging or document services to AI tools.
The introduction of personalised writing features therefore highlights a broader issue in workplace AI: the more context an AI assistant receives, the more personalised its output can become, but the amount of information being made accessible to the system also increases.
OpenAI’s latest move comes amid rapid AI development
The Writing Styles announcement comes shortly after OpenAI introduced GPT-6 Astra, which the company has described as its most powerful AI model yet. The rapid rollout of new capabilities reflects the increasingly competitive AI market, where companies are attempting to improve reasoning, productivity and personalisation at the same time.
Personalised writing is particularly significant because writing is one of the most common uses of generative AI. Users already rely on chatbots to draft emails, rewrite messages, prepare reports and improve grammar.
The new feature attempts to solve one of the biggest criticisms of AI-generated writing: that it can sound repetitive, formulaic or unlike the person supposedly sending it.
What this means for users
For professionals, ChatGPT Work’s Writing Styles could make AI-generated communication more natural and consistent with an individual’s established voice.
A user who frequently writes emails could potentially spend less time rewriting AI-generated drafts simply to make them sound like their own. The system can instead use the user’s existing communication patterns as a reference for future writing.
But the feature also reinforces the importance of checking AI-generated content before sending it. Matching someone’s writing style does not necessarily mean the generated information is accurate, appropriate or suitable for a particular recipient.
Users should also be cautious when connecting accounts containing confidential or sensitive information and should review the permissions available to connected applications.
AI moves closer to personalised digital assistants
OpenAI’s Writing Styles feature represents another step towards AI assistants that understand users on a more personal level.
Rather than treating every prompt as an isolated request, systems are increasingly being designed to learn preferences, working habits and communication patterns. That could make AI considerably more useful for everyday professional tasks.
However, greater personalisation also brings greater responsibility around access to personal and workplace data.
For now, Writing Styles remains a ChatGPT Work feature rather than a standard ChatGPT capability. Whether OpenAI eventually brings the technology to its wider consumer chatbot remains unclear.