Meta AI

The standalone experience needed to feel useful before the conventions for consumer AI products had settled. Our work shaped how people entered a conversation, moved between text and richer inputs, and understood what they could do in a focused web product.

My role concentrated on turning the pitch into a coherent V1: clarifying the product story, resolving the experience with the team, and helping carry it through launch.

Meta.ai did not have the distribution of the mobile surfaces, but it showed stronger retention. That distinction mattered: the web product was not simply another entry point, but a place where people were more likely to return for sustained work.

In parallel, I explored a set of Meta AI productivity concepts and pitched them to Mark Zuckerberg and Chris Cox. The proposals earned buy-in, but the Meta models were not yet capable enough to support the experience we had designed.

Meta later acquired Gizmo and launched Pocket in a related space. Pocket itself was designed by the Gizmo team and others, not by me; the relevance here is the earlier product thesis and the leadership alignment around it.

Across the launch and the exploratory work, the recurring contribution was making an unsettled AI product direction concrete. A strong concept could expose what the model needed to do, what kind of behavior people might value, and whether the organization was ready to invest—even when the immediate answer was not yet to ship it.