Product design question

A prospective customer wants an agentic or RL data solution but can only describe the desired outcome, not the tasks, feedback signals, or delivery constraints. How would you run discovery, separate must-have from nice-to-have requirements, and turn the conversation into a concrete plan for product, operations, and next customer validation?

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

Start a mock interview on this question · Mock interview from a job description

What this question tests

Tests discovery skill for translating a vague outcome into a concrete delivery plan across product, operations, and validation.

How to approach it

  1. Ask the customer to describe the outcome in terms of what success looks like downstream, for example a model behavior they want improved.
  2. Work backward from the outcome to infer candidate tasks and feedback signals, then validate those inferences directly with the customer.
  3. Separate must haves, the minimum data or feedback loop needed to move the metric, from nice to haves, like edge case coverage.
  4. Propose a small scoped pilot task set to validate the approach before committing to full delivery scale.
  5. Translate the validated scope into an operations plan, staffing, quality review process, and a product plan, tooling needed to support annotators.
  6. Set a checkpoint with the customer after the pilot to confirm the data is actually producing the outcome they described.

What a strong answer includes

Common mistakes

Likely follow-up questions

More product design questions

More questions from Scale AI

Learn the skill behind it

Chapters of the AI PM course that teach what this question tests.

Preparing for a specific role?

Book summaries for this kind of question

Browse all 4,000+ questions in the bank