Support Pre-recorded Voicemail Drop for Outbound Calls

We’ve been running outbound sales campaigns on Retell and have noticed an interesting pattern.

Our AI conversations perform well when a prospect answers, but our callback rate from voicemail has been significantly lower than expected.

In previous campaigns, we had much better results when someone from our sales team leaves a real recorded voicemail. Prospects perceived it as a genuine outreach from a person instead of an automated system, which led to noticeably more callbacks.

Currently, Retell allows:

  • AI-generated voicemail prompts
  • Static text messages spoken by the AI voice

However, we haven’t found a way to upload and play an actual pre-recorded audio file (voicemail drop) when voicemail is detected.

It would be extremely valuable if voicemail detection could support something like:

  • Upload a WAV/MP3 recording
  • Play the recording after the voicemail beep
  • Support multiple recordings per campaign or agent
  • Optionally use variables before or after the recording (e.g. the prospect’s first name)

This would allow sales teams to preserve the authenticity of executive outreach while still benefiting from AI-powered outbound calling.

For many outbound campaigns, the voicemail is often the only interaction a prospect has with the company. Hearing a real human voice instead of a synthesized one can make a meaningful difference in callback rates.

Is this currently on the roadmap, or is there another way to accomplish this using Retell?