These are three sample questions from Bryq's AI Fluency for Product Managers assessment, an AI fluency test for hiring product managers, product owners and technical product managers. The full assessment covers three skills: critical eye on AI output, delegating to AI agents and designing reliable AI features. Each question sets out a realistic product scenario involving AI and asks which response fits best, with four options to choose from.
Critical eye on AI output
Intermediate
Question
You manage an online grocery app. You give an AI assistant a churn export with 48,000 customer rows. It reports that churn is highest among weekend shoppers. The finding would reshape next quarter's plan. The assistant's answer reads confidently and includes a chart. Your plan goes to leadership on Thursday.
Select your answer
Ask the assistant how many of the 48,000 rows its analysis actually read.
Give the assistant the export as twelve regional files, one analysis each.
Check that the chart's labels and colors match the churn export's fields.
Ask the assistant to state its confidence in the weekend-shopper finding.
WHY IS THIS THE CORECT ANSWER
An AI assistant might read only part of a large file without saying so, so the first question is how much of the export the finding rests on.
Delegating to AI agents
Intermediate
Question
You manage a project-tracking tool. You ask an AI agent to find out why trial users do not upgrade. Your brief mentions that you suspect the price. The agent writes and sends a ten-question survey to 500 trial users. Its report says 70 percent cite price. Pricing changes are on next week's agenda.
Select your answer
Rerun the survey from a new brief that leaves out your theory about price.
Check that the survey went to a fair mix of new and older trial users.
Ask the agent to send the same survey to another 1,000 trial users this week.
Have the agent interview fifty trial users in depth before any pricing change.
WHY IS THIS THE CORECT ANSWER
An agent writes its questions from the brief it is given, so a brief that names a theory might produce a survey that confirms it.
Designing reliable AI features
Intermediate
Question
You manage a helpdesk product. Your team built an AI prototype that drafts replies to customer emails. In the demo, it answered twenty emails the team had chosen. Every reply read well. Leadership wants to commit the feature for next quarter on that result. Real customer emails vary widely in length and topic.
Select your answer
Run the prototype on a random week of real customer emails before committing.
Ask the team to choose twenty harder emails for a second demo of the prototype first.
Run the prototype on a full year of real emails before any roadmap decision.
Have support leads rate the twenty demo replies for tone before committing.
