Which channels support voice notes?

Inbound voice messages arrive on WhatsApp, Instagram and Messenger. They are downloaded, stored with the conversation and transcribed.

What happens if the audio cannot be transcribed?

The message appears with an audio placeholder and the file is still available to play. Nothing is lost, it just is not searchable as text.

Can the AI answer a voice note?

Yes. The transcript is what the reply agent reads, so a spoken question is answered from the same knowledge base as a typed one.

Is there a size limit?

Yes. Inbound media above the download ceiling is not fetched, which keeps a single large file from blocking a queue.

Can Otobiz make phone calls?

Calling runs on your WhatsApp number, and who is allowed to use it can be limited per person. This one is landing shortly, so the section above carries a marker.

Is there a voice AI agent?

The voice agent answers from the same knowledge base as the messaging agent and follows the same handoff rules. It is marked above because it is landing shortly.

Are transcripts stored with my other data?

Yes. They live in your workspace like every other record, where no other business can reach them.

Can I search old voice notes?

Where a transcript exists, message search covers it, so you can find a conversation by what was said in it.

Customer conversations

Voice notes become text.

Customers who would rather talk than type get answered anyway.

Voice messages are the default for many customers, and they break most support tooling because a queue full of audio cannot be searched, scanned or answered quickly. Transcription converts an inbound voice note into text in the same thread, so it can be read at a glance, searched later, and answered by the same grounded reply path as a typed message.

Inbound voice notes transcribed into the conversation
Transcripts searchable with the rest of your messages
The reply agent drafts from the customer's actual words
A file too large to handle is called out, not silently dropped

How it works.

01
Customer records a note
Someone sends a voice message on WhatsApp, Instagram or Messenger instead of typing, which is common on mobile and in noisy environments.
02
Audio is fetched
The file is downloaded from the provider within a size limit and stored against the conversation like any other attachment.
03
Speech becomes text
The audio is transcribed and the text is attached to the message, so your team reads it instead of pressing play.
04
It is answered like a message
The transcript feeds the same grounded reply path, so a spoken question about an order gets the same answer a typed one would.

Audio your team can scan

A queue of voice notes is slow to triage because every one costs its full length to hear. Text makes it scannable in the same second as any other message.

Transcript shown with the message in the thread
Included in message search
The original audio still available to play

Handled inside the size limits

A file too large for the channel is refused with a reason instead of arriving broken, and audio is converted to a format that plays where it needs to.

Download ceiling on inbound media
Format conversion so the audio plays reliably
A clear placeholder when transcription is unavailable

The same rules as text

A transcribed message is a message. It carries the same consent state, the same conversation window and the same audit trail.

Grounded reply drafting from the transcript
Consent and window rules unchanged
Sent and logged the same way as a typed message

Calls on WhatsApp^soon

Business calling on the WhatsApp number, so a conversation that needs a voice call does not have to leave the channel it started in.

Inbound and outbound calls on the connected number
Call records attached to the customer's thread
Access to calling controlled per person

A voice agent that answers^soon

An automated voice answer for calls, grounded in the same knowledge base the messaging agent uses, with the same approval boundaries.

Spoken answers from the same source material
Handoff to a person on the same rules
Recording and transcript kept with the conversation

What it does not do.

A voice note only transcribes as well as it was recorded. Very noisy audio still needs a person to listen.
Calling runs on your WhatsApp number. It is not a contact centre phone system with queues and hold music.
Outbound voice messages from your team are not part of the transcription path.

Common questions

Which channels support voice notes?: Inbound voice messages arrive on WhatsApp, Instagram and Messenger. They are downloaded, stored with the conversation and transcribed.
What happens if the audio cannot be transcribed?: The message appears with an audio placeholder and the file is still available to play. Nothing is lost, it just is not searchable as text.
Can the AI answer a voice note?: Yes. The transcript is what the reply agent reads, so a spoken question is answered from the same knowledge base as a typed one.
Is there a size limit?: Yes. Inbound media above the download ceiling is not fetched, which keeps a single large file from blocking a queue.
Can Otobiz make phone calls?: Calling runs on your WhatsApp number, and who is allowed to use it can be limited per person. This one is landing shortly, so the section above carries a marker.
Is there a voice AI agent?: The voice agent answers from the same knowledge base as the messaging agent and follows the same handoff rules. It is marked above because it is landing shortly.
Are transcripts stored with my other data?: Yes. They live in your workspace like every other record, where no other business can reach them.
Can I search old voice notes?: Where a transcript exists, message search covers it, so you can find a conversation by what was said in it.

Every question in one place, answered once.