---
title: "The Interview Trick"
description: "Voice mode flips authoring into answering: have Claude interview you."
order: 7
duration: "10 min"
chapter: "02-claude-in-your-pocket"
type: lesson
---

## A conversation while you walk

Dictation is one-way: you talk, Claude transcribes, you get a document. Voice mode is different — it is a two-way spoken conversation. You talk, Claude talks back, and you never look at the screen.

The two modes have different jobs:

| Mode | Shape | Use It For |
|------|-------|-----------|
| **Dictation** | You talk → text appears in the message box | Capture. Getting knowledge out of your head verbatim. You want YOUR words on the record. |
| **Voice mode** | Spoken back-and-forth | Thinking. Working through a problem, rehearsing a conversation, being interviewed. You want dialogue. |

The interview pattern is where voice mode earns its place in this course. Explaining what you know is hard because you do not know what you know — expertise is invisible to the expert. But answering questions is easy. So flip it: have Claude interview *you*.

Sarah tries it on her drive between the two stores. She starts voice mode and says:

> "I need to document how we handle supplier problems — damaged goods, wrong deliveries, pricing disputes. Interview me about it. Ask me one question at a time, and dig into the details like you're a consultant writing our operations manual."

Claude asks: "Let's start with damaged goods. Walk me through the last time a delivery arrived damaged — what happened, step by step?" Sarah answers. Claude follows up: "You said you photograph the damage *before* signing the delivery receipt — is that a rule everyone knows, or just your habit?" Twenty minutes of driving later, Claude has extracted a damage-claims process, an escalation ladder for pricing disputes, and two rules Sarah had never said out loud in twelve years of running the store. At a red light, she says: "Summarize everything as a process document." Done before she parks.

<div class="exercise">
  <div class="callout-label">Try This</div>
  <p>Start voice mode during otherwise-dead time — a commute, a walk, doing dishes. Say: "Interview me about [a process you own at work]. One question at a time. Dig into edge cases and exceptions like you're writing our operations manual." Answer out loud for ten minutes. Then ask for a written summary.</p>
  <p>Compare it to what you would have produced staring at a blank document titled "Our Process." The interview version is almost always richer — because answering is easier than authoring.</p>
</div>

## Check your understanding

Answer in your own words — write it down before opening the key. Your tutor grades against the criteria and generates fresh variants on retries.

**Q1.** You need to document a process you own, but staring at a blank page produces nothing. Write the exact voice-mode prompt you would use, and explain why this pattern works when authoring fails.

**Q2.** Dictation and voice mode are different tools. Give one job each is right for, and explain what goes wrong if you swap them.

<details>
<summary>Answer key — attempt every question first</summary>

## Answer key

### Q1

**Model answer:** Something like: "Interview me about how we handle [process]. Ask one question at a time, and dig into edge cases and exceptions like a consultant writing our operations manual." It works because answering questions is cognitively easy while authoring is hard — expertise is invisible to the expert until a question forces it out.

**Pass criteria:** prompt includes interviewing + one-question-at-a-time + digging into specifics; explanation contrasts answering vs authoring

### Q2

**Model answer:** Dictation is for capture — getting YOUR words on the record verbatim, e.g. recording vendor rules while walking the aisles. Voice mode is for dialogue — thinking through a problem or being interviewed on a commute. Swapped: voice mode paraphrases when you wanted your exact rules; dictation cannot ask the follow-up question that extracts what you forgot to say.

**Pass criteria:** assigns capture to dictation and dialogue/thinking to voice mode; failure modes reference one-way vs two-way nature

</details>


**Next:** [Point Your Camera](/mastering-claude/02-claude-in-your-pocket/08-point-your-camera/)