Most AI-generated text has a tell. The sentences are too balanced, the vocabulary too clean, the sign-offs too formal. OpenAI is now directly targeting that problem. ChatGPT has announced a new personalization feature that analyzes your existing writing across apps like Gmail, Google Drive, Slack, and SharePoint, and then carries those patterns into whatever you draft next inside ChatGPT.
The feature, called Writing Style, sits inside ChatGPT’s Personalisation settings. Once enabled, the system reads through your emails, messages, and files to build a profile of how you actually write. That includes vocabulary choices, sentence structure, punctuation habits, preferred sign-offs, and even capitalization quirks. The goal is that when ChatGPT writes a draft for you, it should sound like a slightly faster version of you, not like a generic language model on autopilot.
To access it, go to ChatGPT on the web, open Settings, click Personalisation, and select Writing Style. The feature applies across both the web and mobile versions and is currently available to paid plan users with access to ChatGPT Work.
This matters because generic AI output is one of the biggest friction points for professionals who rely on these tools daily. There’s a significant difference between AI that produces competent text and AI that produces text you’d actually send without rewriting half of it. Style personalization closes that gap, at least in theory.
The broader trend here is worth paying attention to. Google’s Gemini already offers some level of tone customization in Google Docs. Microsoft Copilot, deeply embedded in the Microsoft 365 ecosystem, also learns from user behavior across Teams and Outlook. OpenAI is now pushing into similar territory, but by pulling in data from multiple platforms rather than staying confined to one product suite. That cross-app approach is what separates this from competitors at the moment.
Under the hood, the system uses natural language processing to break down text into components:
- Word choice and vocabulary preferences
- Sentence length and structural patterns
- Emotional tone and formality level
- Punctuation and capitalization habits
- Recurring phrases and sign-off styles
Together, these form what researchers sometimes call a linguistic fingerprint, a consistent set of patterns that shows up across different types of writing from the same person.
For founders and developers building on top of ChatGPT, this signals where OpenAI sees its competitive edge: not raw model capability, but stickiness. The more your work lives inside ChatGPT, the harder it becomes to switch. That’s a deliberate product strategy, and this feature is a clear part of it.
