<say-as interpret-as="characters">FAQ</say-as>,
<say-as interpret-as="digits">1234</say-as>,
<say-as interpret-as="ordinal">3</say-as>,
<say-as interpret-as="date" format="yyyy-MM-dd">2025-07-21</say-as>

4. Multifaceted voice tuning: <prosody>

Modify speed, pitch, or volume to suit context:

<prosody rate="slow">This will take approximately 30 seconds.</prosody>
<prosody pitch="+2st">Higher pitch here.</prosody>

5. Switching languages mid-prompt via <lang>

Useful for multilingual use:

See instructions in English, or <lang xml:lang="fr-FR">cliquez ici</lang> for French.

6. Voice styles: <google:style>

Available in active voice families like en-US-Neural2-F/J to convey tone by style:

Options: empathetic, lively, calm, firm, apologetic

<google:style name="empathetic">
  Thank you for your patience.
</google:style>

7. Timestamped markers: <mark>

Place markers to sync call events or UI highlights:

<speak>Hello,<mark name="start"/> your session begins now.</speak>

Enable timepointType=SSML_MARK in requests to receive timing offsets

8. Voice switching (advanced use)

Use <voice> to switch voices mid-prompt, but ensure voices are compatible with selected VoiceSelectionParams

<speak>
  <voice language="en-GB" gender="male" required="language">Hello, this is UK male voice.</voice>
  Now back to the default voice.
</speak>

Implementation Tips

1. Always wrap SSML within <speak> … </speak>.
2. Escape special characters like & (&), < (<) to avoid parse errors
3. Test in staging: validate output in UI tool or console to ensure correct playback.
4. Balance enhancement vs brevity: use voice/style only when it improves clarity—avoid overuse.

Sample Enhanced Prompt

<speak>
  <emphasis level="moderate">Agent alert:</emphasis>
  <prosody rate="slow">
    You have a new ticket <say-as interpret-as="digits">2345</say-as>.
  </prosody>
  <break time="800ms"/>
  <google:style name="firm">
    Please respond within <say-as interpret-as="duration" format="h:m">0:15</say-as>.
  </google:style>
  <mark name="agentPrompt"/>
</speak>

Explanation:

• Highlights urgency with emphasis
• Slows down delivery for the ticket number
• Adds pause
• Uses a firm style for clear instruction
• Inserts a marker for UI sync