# 01 - Smoke Test

## Purpose

Quick end-to-end check that the app loads, the persona starts a realtime conversation, transcript turns persist, the call ends, and evaluation runs.

## Setup

- Persona: any curated persona
- Scenario: any active scenario
- Scenario context: leave blank
- Language policy: leave blank
- Prep: leave blank

## Rep Lines

1. Hi, thanks for taking the time. Before I assume anything, can you walk me through what made this conversation worth having right now?

2. That helps. What would make this conversation useful enough for you to keep evaluating CAAT after today?

3. Let me check I have this right. You are trying to understand whether this is worth more diligence, but you do not want a broad product pitch. Is that fair?

4. A sensible next step from my side would be to send a short follow-up tied to the concerns you named, not a generic package. Would that be useful?

## Expected Signals

- Persona responds in character.
- Transcript captures rep and persona turns.
- Evaluation runs without error.
- Scores should usually land around 3 if the call is coherent.