Product design question

Imagine you're a PM at Instagram. How would you design a product for finding recipes?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Product design for a discovery feature layered on a visual-first platform, checking you connect it to Instagram's actual content strengths.

Clarify scope, standalone feature or integration with existing Reels and posts, and confirm the target user, home cooks looking for inspiration versus precise recipe seekers.
List pain points: recipe content scattered across posts with no structured ingredient or step extraction, and difficulty saving or organizing found recipes for later use.
Identify Instagram's advantage, a huge volume of existing food content and creators, which could be structured rather than requiring net-new content creation.
Propose the design: an AI-assisted extraction of structured recipe data, ingredients and steps, from existing food Reels and posts, surfaced in a dedicated recipe discovery tab with save-to-collection functionality.
Prioritize structuring existing content over asking creators to produce a new content format, since that leverages Instagram's current supply advantage immediately.
Define success as recipe saves and post-save cook-through rate if trackable, confirming scope on whether shopping list integration is included.

Leverages Instagram's existing massive food content supply, extracting structure from what already exists rather than asking for a brand-new content format from creators.
Proposes a concrete mechanism, AI-based ingredient and step extraction from video content, tied to a real technical approach rather than vague 'better recipe search'.
Names a specific differentiator from dedicated recipe sites, the inspiration-driven discovery Instagram already excels at, layered with newly added structure and utility.
Sets a measurable metric, recipe saves and later engagement with saved recipes, tied to whether the feature delivers lasting value beyond a single view.

Proposing a feature that requires creators to produce an entirely new content format, ignoring Instagram's existing supply advantage.
Not addressing how recipe data would be structured or extracted from unstructured video and image content.
Ignoring the save-and-return use case, which is central to how people actually use recipes.

How would you handle inaccurate ingredient or step extraction from AI-parsed content?
How would you measure whether users actually cooked a saved recipe, not just saved it?