Skip to main content
Anywhere your flow speaks to a caller, you choose how that message is made: record it yourself, upload a file, or write the text and let Allo read it out.

​
The three sources

On any step that talks, select Edit on the audio row and pick one:
Uploads must be MP3 or WAV, up to 5MB. Other audio formats can look like they saved and then play as silence to your callers. If you have an m4a file, convert it before uploading.

​
One source at a time

A step plays one thing. If a step has a recording and you then write text for it, the text replaces the recording, and Allo warns you before it does: “This step plays a recording. The text you write here replaces it.” Menu option names are the exception. There, the name you type is also the label on the canvas, so the name and the recording live side by side: the name is what you read on screen, the recording is what the caller hears. The row shows Named by your recording, and Remove recording takes you back to the spoken name.

​
Where audio shows up

  • Phone menu greeting, played before the options.
  • Menu option names, the “For Sales, press 1” line for each key.
  • Play announcement, the message played mid-flow.
  • Voicemail greeting, played before the caller records.

​
Writing text that reads well

Allo reads your text exactly as written, so write it the way you would say it.
  • Spell out anything you want heard as words. “24/7” is read back as digits and a slash.
  • Give phone numbers room to breathe: “zero one, four four, five five” reads better than one long string.
  • Punctuation sets the pace. A full stop is a real pause.
  • Listen to it once with the player on the row before you save.

​
Announcements need their text

A Play announcement step will not save without written text, even when a recording is what actually plays. Write the message first, then add the recording if you want a human voice on it. The text stays as the label for that step.

​
Recordings and barge-in

With a text-only menu, a caller can press their key at any point while the menu is being read. As soon as any part of a menu uses a recording, that changes: callers can only press once the last option starts playing. If your regulars are used to pressing 1 straight away, keep that menu on text to speech.

Build a phone menu

Greeting, options and fallbacks

Call flow troubleshooting

Callers hear silence, or hear the menu twice