Marissa Mayer’s startup has unveiled Dazzle, a personal AI assistant that builds its understanding of a user from the photos stored on their phone rather than primarily from email, calendars or shopping history. The company previously raised an $8 million seed round.
Users can interact through an app or text messages. Dazzle can scan recent images to extract immediate details, such as placing an event from a photographed flyer onto a calendar. It also claims to infer longer-term information including hobbies, travel, food and style preferences, and the people with whom a user spends time. The approach draws on technology from Mayer’s earlier photo-sharing product, Shine, which shut down after failing to gain broad use.
A camera roll can reveal locations, relationships, documents and private moments, making permissions central to the product rather than a secondary setting. Dazzle’s usefulness will depend on whether people can see, correct and limit what it infers, and whether sensitive analysis stays protected. The demo shows a distinct source of personal context, not yet proof that users will trust it.