Metrics question

How would you measure user trust in an autonomous agent over time?

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

Metrics design for a hard-to-quantify construct, trust, tracked over time rather than as a single snapshot.

How to approach it

  1. Define trust behaviorally: users delegate progressively more consequential or higher-stakes tasks to the agent over time without needing to double-check its work.
  2. Track a delegation-depth metric: the average risk or complexity level of tasks a given user assigns to the agent, watched for growth over their tenure as a user.
  3. Track a verification-behavior metric: how often users manually double-check or redo the agent's output, expecting this to decline over time as trust builds.
  4. Track retention and expansion as lagging trust indicators, since users who do not trust an agent will simply stop using it or limit it to trivial tasks.
  5. Add a direct survey signal periodically, asking users to rate their confidence in the agent for specific task types, to validate the behavioral proxies.
  6. Define success as an upward trend in delegation depth combined with a downward trend in manual verification, tracked per user cohort over their first several months.

What a strong answer includes

Common mistakes

Likely follow-up questions

More metrics questions

More questions from Manus

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