Product Sense: Designing an AI Conversational Assistant Inside Instagram

You are a product manager at Meta and have been asked to design an AI chatbot for Instagram, the company's photo and video sharing app. Decide whom it serves and what it does, describe the experience, and explain how you would measure whether it is succeeding. ```hint Pick the job before the bot A chatbot is a format, not a product. Choose one Instagram user group and one recurring task where a conversation beats the taps and screens they use today, then design for that. ``` ### Clarifying Questions - Who should the chatbot serve: people browsing and messaging on Instagram, creators, or businesses that sell and handle customer messages there? - Where should it live: inside direct messages, in search and Explore, in the camera and editing tools, or as a separate surface? - Is the main goal engagement, productivity for creators and businesses, or revenue? - Should the design assume Meta's existing general-purpose AI assistant is already available in the app, so this is a specific new use case, or start from nothing? ### What a Strong Answer Covers - A clear goal tied to Instagram's role in Meta's business, and a justified choice of one user segment and one use case - Why a conversational interface beats the existing Instagram flows for that job - A concrete experience: entry points, what the bot can and cannot do, hand-off to a human, and clear AI labeling - Risks specific to an AI on Instagram: wrong or invented answers, impersonation and scams, teen safety, and the privacy of messages - A success metric with guardrails that catch harm to human interaction, trust or safety ### Follow-up Questions - How would you keep the chatbot from replacing the person-to-person conversations that make Instagram valuable? - A teenager starts using the chatbot every night for emotional support. What should the product do? - What result after a three-month test would make you stop the project?