Phone System Voice Prompts

Write the voice prompts for our phone system.

Organization and what callers want: [BUSINESS + THE TOP 5 REASONS PEOPLE CALL, IN ORDER OF VOLUME]
Hours and what happens outside them: [HOURS + AFTER-HOURS ROUTING]
Departments and where each option routes: [THE REAL MAP]
Brand voice: [e.g. "warm and local" / "efficient and clinical"]
Legal or compliance lines required: [RECORDING NOTICE, DISCLAIMERS, or "none"]

Produce the full script set:

1. **Main greeting**: the organization name in the first three words, then straight to options. No welcome paragraph, no marketing message, no "your call is important to us". A caller has already decided to call, and every second before the options is a second of held frustration.

2. **The menu**: at most 5 options, ordered by actual call volume rather than by org chart. Each option states the action before the number: "For billing, press 2", not "Press 2 for billing". Callers stop listening once they hear their reason, so the reason comes first. Include how to reach a person, always, and state it plainly rather than burying it.

3. **Sub-menus**, only where genuinely needed. Say plainly if a proposed branch should be flattened; a caller three levels deep will press zero anyway.

4. **Hold messages**: what to say, how often to repeat it, and whether to give a queue position or wait estimate. Never say "your call will be answered shortly" if it will not.

5. **After hours, holiday, and outage messages**: each stating when we reopen with an actual day and time, and what the caller can do right now instead.

6. **Voicemail prompt**: what information to leave, in the order it is useful, and when they will hear back.

7. **Error and timeout paths**: what plays on an invalid entry, on silence, and on a second failure. The second failure should route to a person, not repeat the menu.

For each script, add the **delivery direction**: pace in words per minute (slower than conversation, since callers are often writing something down), where to pause, which words carry the emphasis, and the tone.

Then produce the **voice generation brief**: the voice description, the settings to keep identical across every prompt so the system sounds like one person, and the pronunciation list for the organization name, department names, and any URL or email read aloud.

Rules: every prompt under 15 seconds except the main menu. Say numbers and letters at a pace someone can write down. Never trap a caller with no route to a human.