interview-me

You can clarify underspecified requirements by conducting a one-question-at-a-time interview to extract true user intent, generating hypotheses, and scoring confidence until it reaches approximately 95% for explicit confirmation.

Start by stating a hypothesis with an attached confidence score, then ask one clarifying question at a time, updating your confidence level until it reaches roughly 95% to trigger a structured restate for explicit confirmation.

Use hypothesis-driven questioning when requests are underspecified or when you explicitly invoke interrogation to surface hidden requirements, ensuring you attach a confidence score to your initial guess to identify and address information gaps.

The best way to reach high confidence is through sequential one-question prompts combined with continuous confidence scoring, stopping the interview process only when your confidence level hits around 95%.

Yes, the approach strictly enforces a one-question-at-a-time flow to prevent batching, deliberately reducing assumption-driven design by guiding a focused interview that surfaces the true underlying goal incrementally.