How do you proceed when requirements are unclear?

Question

Accepted Answer

Ambiguity is the norm in real work. They want to see you **reduce uncertainty actively** rather than freeze or guess silently and build the wrong thing.

## How to approach it

```text
1. Ask the questions that change what you'd build
2. Find the goal behind the request (the "why")
3. State your assumptions in writing and confirm them
4. Build a thin slice, demo early, adjust fast
```

## Worked example

```text
S: A ticket said "add a dashboard" with no detail on metrics or users.
T: Building blind risked weeks of wasted work.
A: I asked who'd use it and what decision it drove, wrote down three assumptions,
   and built a clickable prototype of the top two metrics first.
R: The demo surfaced that they actually needed exports, not charts. We pivoted
   after two days instead of two weeks.
```

## Good vs weak

```text
✓ Clarifies the "why," states assumptions, demos early
✗ Building the full thing on a guess
✗ Refusing to start until everything is spec'd
```

## Why it matters

Clear specs are rare; the engineers who thrive are the ones who can move through fog.

Finding the underlying goal prevents you from building exactly what was asked but not what was needed.

Demoing a thin slice early turns expensive guesses into cheap, fast corrections.