Marissa Mayer pivots to photo-based AI with Dazzle

Dazzle, which recently closed an $8 million seed round, differentiates itself from competitors like Meta’s Muse by treating visual history as the primary data source. Mayer argues that a user’s photo collection provides a deeper layer of insight into personal style, family interests, and past experiences. The assistant functions by scanning images to perform immediate tasks, such as extracting calendar details from event flyers or suggesting personalized vacation spots based on previous travel history.

Technological roots for the product trace back to Mayer’s previous startup, Sunshine, and its photo-sharing tool, Shine. Although that project eventually shuttered, Mayer notes that the intellectual property proved foundational for Dazzle’s current capabilities. During testing, the assistant successfully identified niche activity interests, such as escape rooms and kayaking tours, though it occasionally struggled with specific details like a child’s existing skill level.

Privacy remains a central pillar of the company’s pitch. Mayer suggests that users are likely to feel more comfortable granting AI access to photos than to sensitive email or message archives. Dazzle claims to discard any information flagged as private, positioning itself as a more personal alternative to conventional digital assistants.