Setting 12-Month Goals and a North-Star Metric for Meta AI Six Months After Launch

Meta's AI assistant went live six months ago, and the company wants it to become a core experience. As its product manager, how would you set goals for the next 12 months? ```hint Define core before counting Decide what it would look like for the assistant to be part of people's routine rather than something they tried once, then work backward to measurable goals and to the order in which you would pursue them over the year. ``` ### Clarifying Questions - Which surfaces are in scope: the assistant inside WhatsApp, Messenger, Instagram and Facebook, a standalone app, smart glasses, or all of them? - What does leadership mean by core: a daily habit for a large share of Meta's users, a driver of engagement in the existing apps, or a future business line? - What do we know from the first six months about reach, repeat usage, the most common use cases, and quality or safety incidents? - Are these goals for the assistant as a whole, or does each surface team own its own targets? ### What a Strong Answer Covers - A clear definition of core experience, turned into a north-star metric that rewards repeated, successful use rather than raw reach - Input metrics the team can directly move, such as activation, answer quality, use-case breadth and latency, and how they add up to the north star - Guardrails for safety, accuracy, privacy, compute cost and the health of Meta's existing apps - How targets are derived from the six-month baseline, sequenced across the year and split by surface and market - How progress is reviewed and when the goals themselves would be revised ### Follow-up Questions - Monthly users jumped after the assistant was added to an app's search bar, but repeat usage fell. Was that quarter a success? - How would you set a target for answer quality when open-ended conversations have no single correct answer? - Leadership wants one number for the company dashboard. Which do you choose, and what does it hide?