From reactive to proactive
Until now, Apple Intelligence responded to commands. With iOS 27, the system acquires a proactive dimension: Tom's Guide and Digit.in document that the intelligence can now read the content of messages, emails, and other on-screen elements to suggest relevant actions before the user requests them. If someone texts you about an upcoming appointment, Apple Intelligence can suggest automatically creating a calendar event. If a friend asks for photos from a recent trip, the system can surface relevant images and facilitate sharing.
Integration with Messages, Mail, and Calendar
Digit.in describes the mechanics in detail: Messages can propose actions based on conversation context, including reminders or searches in the photo library; Mail adopts a Smart Reply that adapts to the user's writing style; Calendar can create or edit events simply by interpreting a natural language description. Tom's Guide notes that the system can locate photos using details mentioned in conversations — people, places, keywords — and surface relevant information from messages and emails when it becomes useful.
The trust question
This type of integration raises legitimate privacy questions: a system that constantly reads conversations and mail to suggest actions is powerful, but requires a high level of trust. Apple reiterated during the keynote that processing happens on-device or through Private Cloud Compute without storing personal data. The statement is consistent with the company's past positions, but practical verification — especially for cloud-dependent features — remains to be tested during beta.