Meta AI Inside Instagram DMs: Pick a Goal, Prioritize Features, Set Metrics

Quick Overview

A product manager prompt about improving Meta AI, Meta's assistant, inside Instagram direct messages. It tests choosing a goal beyond raw chat volume, picking a messaging segment, prioritizing assistant features that use Instagram content and group chats, and defining success metrics with privacy and safety guardrails.

Meta AI Inside Instagram DMs: Pick a Goal, Prioritize Features, Set Metrics

Company: Meta

Role: Product Manager

Category: Product Design & Strategy

Difficulty: medium

Interview Round: Other

You are a product manager at Meta working on Meta AI, the company's AI assistant, where it shows up inside Instagram direct messages. How would you make that assistant experience better? Walk through the goal you would set, the users you would focus on and the problems they have, the improvements you would build and in what order, and how you would know they worked. ```hint Use what only Instagram has General-purpose chatbots already exist in many apps. Ask what an assistant can do inside Instagram messaging that it cannot do anywhere else, given the Reels, posts, friends and group chats around it, and let that shape the goal and the problems you pick. ``` ### Clarifying Questions - Why improve it now: is the concern that few people try the assistant, that people try it once and do not come back, or a strategic push toward assistants that complete tasks for people? - Does the scope cover only the one-to-one thread with the assistant, or also bringing it into group chats and into posts or Reels that friends share in messages? - Are teen accounts in scope, given the extra safety and privacy protections they require? - Is success judged by engagement with the assistant itself, or by Instagram-wide outcomes such as messaging activity and retention? ### What a Strong Answer Covers - A goal that ties the assistant to how people actually use Instagram messaging (sharing and discussing content with friends), with a reason it beats raw message volume - A chosen segment of Instagram messaging users and pain points specific to how they use DMs, Reels and group chats - Assistant improvements prioritized by impact, confidence and effort, including how people discover them - A success metric that counts useful conversations rather than message totals, plus guardrails for person-to-person messaging, answer quality and inference cost - Risks specific to an AI inside private messages: consent in group chats, wrong or unsafe answers, and protections for younger users ### Follow-up Questions - Messages to the assistant go up, but messages between friends in the test group drop slightly. Do you ship? - How would you let someone bring the assistant into a group chat without the other members feeling watched? - How would you test a group-chat feature when members of the same chat could land in different experiment arms?

Overview: A product manager prompt about improving Meta AI, Meta's assistant, inside Instagram direct messages. It tests choosing a goal beyond raw chat volume, picking a messaging segment, prioritizing assistant features that use Instagram content and group chats, and defining success metrics with privacy and safety guardrails.

Meta logo
Meta
Aug 21, 2026
mediumProduct ManagerOtherProduct Design & Strategy
0
0

You are a product manager at Meta working on Meta AI, the company's AI assistant, where it shows up inside Instagram direct messages. How would you make that assistant experience better? Walk through the goal you would set, the users you would focus on and the problems they have, the improvements you would build and in what order, and how you would know they worked.

Clarifying Questions Guidance

  • Why improve it now: is the concern that few people try the assistant, that people try it once and do not come back, or a strategic push toward assistants that complete tasks for people?
  • Does the scope cover only the one-to-one thread with the assistant, or also bringing it into group chats and into posts or Reels that friends share in messages?
  • Are teen accounts in scope, given the extra safety and privacy protections they require?
  • Is success judged by engagement with the assistant itself, or by Instagram-wide outcomes such as messaging activity and retention?

What a Strong Answer Covers Guidance

  • A goal that ties the assistant to how people actually use Instagram messaging (sharing and discussing content with friends), with a reason it beats raw message volume
  • A chosen segment of Instagram messaging users and pain points specific to how they use DMs, Reels and group chats
  • Assistant improvements prioritized by impact, confidence and effort, including how people discover them
  • A success metric that counts useful conversations rather than message totals, plus guardrails for person-to-person messaging, answer quality and inference cost
  • Risks specific to an AI inside private messages: consent in group chats, wrong or unsafe answers, and protections for younger users

Follow-up Questions Guidance

  • Messages to the assistant go up, but messages between friends in the test group drop slightly. Do you ship?
  • How would you let someone bring the assistant into a group chat without the other members feeling watched?
  • How would you test a group-chat feature when members of the same chat could land in different experiment arms?