Strategy question

Imagine the research team has resources for only one of two model directions this quarter: improving audio quality/coherence or improving controllability and prompt following. How would you help Suno make that tradeoff using user insight, business impact, and evidence from model evaluations?

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

Tests using user insight, business impact, and model-eval evidence to make a hard research-resource tradeoff between two improvement directions.

Gather user insight on both directions: feedback or support signals indicating whether users are more often frustrated by low audio quality or by the model not following their prompt.
Gather business impact signals: which direction more directly affects retention or upgrade conversion, since both matter differently to different user segments.
Gather model-eval evidence: current benchmark gaps in each direction, to see which has more near-term room for meaningful improvement versus diminishing returns.
Weigh audience segmentation: casual users may care more about baseline quality, while power or Suno Studio users likely care more about controllability.
Make the recommendation explicit, naming which direction and why, rather than presenting both as equally valid to avoid a hard call.
Propose a lightweight validation step, like a targeted user survey or a quick internal eval, before fully committing the quarter's resources.

Uses all three evidence types, user insight, business impact, and eval data, together rather than picking based on one alone.
Ties the recommendation to audience segmentation, recognizing casual and power users may value the two directions differently.
Makes an explicit recommendation instead of presenting a balanced non-answer to a genuinely forced tradeoff.

Avoiding a real recommendation by saying both directions matter equally.
Ignoring segment differences between casual users and power users in weighing the tradeoff.

What would change your recommendation if power-user retention data came in differently?
How would you validate the chosen direction before committing the full quarter?