Skip to main content
Enquiry
Techmits
Automation & AI

Voice Note Transcription & Spoken AI Replies

Incoming voice notes are automatically transcribed to text, and the AI bot can optionally reply with a spoken voice note of its own.

Last updated 2026-08-14

When a customer sends a voice note instead of typing, WhatsApp Cloud automatically converts it to text so your team can read it at a glance. If the AI Auto-Reply Bot is enabled, it can optionally speak its reply back as a voice note too, instead of just text.

Automatic voice note transcription

Any voice note a customer sends is automatically transcribed to text and shown underneath the audio player in the conversation thread.

This happens automatically for every genuine voice note — there's nothing to turn on.
A forwarded audio file (like a shared song) is treated differently from a tap-to-record voice note and isn't auto-transcribed or treated as a question to answer.
If the AI bot is enabled, it reads the transcript, not just the audio file, so it can respond to what the customer actually said — including in a voice note sent a couple of messages back.

Spoken AI replies

Instead of just a text reply, the AI bot can optionally speak its answer back as a voice note.

  1. Go to Automation from the left menu.
  2. Turn on "Reply to voice notes with voice".
  3. Save.
Off by default — turning it on is a deliberate choice, since a spoken reply is a bigger commitment than a text one.
There's currently one fixed voice for every account — choosing between different voices isn't available yet.
If generating the spoken reply fails for any reason, the bot automatically falls back to a normal text reply rather than sending nothing.
A spoken AI reply is labeled the same way any AI reply is — see AI Auto-Reply Bot.

Frequently Asked Questions

Do I need to do anything for voice note transcription to work?

No — transcription of incoming voice notes happens automatically and doesn't need to be turned on.

Can I choose a different voice for spoken AI replies?

Not yet — there's one fixed voice used for every account that enables spoken replies. Voice selection is on the roadmap.

What happens if the voice reply fails to generate?

The customer still gets a reply — WhatsApp Cloud automatically falls back to a normal text message if anything goes wrong creating the voice note, so the customer is never left without a response.

Does a forwarded song or audio clip trigger the AI bot?

No — only a genuine tap-to-record voice note is transcribed and treated as a message to respond to. A forwarded audio file is left alone.

Ready to Get Started?

See it for yourself

Create a free account and connect your WhatsApp Business number in minutes.

Get Started