Marissa Mayer Is Back with Dazzle, the New AI Assistant That Learns From Your Camera Roll

She explains that Dazzle asks for permission to access your camera roll, then uses historical photos to learn tastes, lifestyle, and important people/places, while using new photos to infer intent and trigger actions like buying, booking, calendaring, or note capture.

She says work at her prior startup revealed only about 60% of photos were “shareworthy,” and the remaining ~40% were typically reminders, coordination, or “get it done” items—leading to the concept of an assistant that activates when a photo represents a task.

Marissa Mayer argues photos provide high-density personal context without extra work from the user, citing examples like recognizing books on shelves in the background of everyday photos and identifying appliances by make/model from incidental capture.

She describes Dazzle’s goal of making the camera button a “get it done” button, turning captured items into a “get-it-done list” that can execute via AI agents rather than just listing tasks.

She says Dazzle emphasizes fast onboarding for non-technical consumers, claiming users typically complete their first task within 45 seconds and without the setup burden of connectors.