Alright, let's cut through the marketing. Everyone says "brand voice" but few actually deliver a usable, concrete guideline document for their team. I see people dumping a bunch of .mp3 files from WellSaid into a shared drive and calling it a day. That's useless.
Here's a step-by-step method to actually build something actionable. This assumes you've already generated your core samples in WellSaid Labs.
**First, Structure Your Document:**
Your guideline doc needs these sections:
* **Voice Attributes:** The abstract stuff (e.g., "Authoritative but approachable").
* **Sample Library:** The concrete proof. This is the most important part.
* **Technical & Usage Rules:** Where the rubber meets the road.
**The Critical Step: Annotating Your Samples**
Don't just link to the audio. For each WellSaid sample, create a table entry. This forces you to define *why* it works.
| Sample Name | WellSaid Voice Used | Scenario/Purpose | Transcript Snippet | Why It Works / Notes |
| :--- | :--- | :--- | :--- | :--- |
| `onboarding_welcome.mp3` | `Sonia` | New user welcome | "Welcome aboard. Let's get you set up..." | Warm, paced delivery. Slight upward inflection on key verbs. |
| `error_alert_critical.mp3` | `Dax` | System outage alert | "Critical service interruption detected..." | Urgent but controlled tone. No panic. Measured pace. |
| `feature_update_excited.mp3` | `Maya` | Product launch announcement | "We're thrilled to introduce..." | Higher energy, faster pace, but still clear enunciation. |
**Cost Consideration (because I have to):**
Generating 20-30 samples for this exercise will have a real cost. If your team plans to iterate constantly, your WellSaid bill will look like this:
```bash
# Hypothetical but realistic monthly pattern
Base Voices: 20 samples * $0.1/min = $2.00
Custom Voice Training: Fixed cost (one-time)
Additional Usage: ~500 mins/month * $0.1 = $50.00
```
**You need to track this as a project expense.** Set a budget alert in WellSaid for $100 if you don't want surprises.
The final deliverable is a PDF with the tables and embedded audio links, plus a clear section on which voice to use for "Tier 1 Support vs. Marketing Explainer." Without that, you've just made a soundboard, not a guideline.
show me the bill
show me the bill