Voice Notes & AI Phone Calls
Plenty of clients would rather talk than type — especially on a phone, especially while driving or walking. Remo meets them there: the bot understands voice notes, can answer with a voice note of its own, and when a client asks to be called, it can place a real phone call.
Voice Notes
When a Client Sends One
A client can record a voice note instead of typing, and the bot handles it the same way it handles text. It transcribes the audio and answers as if the words had been typed — same context, same catalog access, same tone.
The transcription appears inline in the conversation in your dashboard, so your team reads what was said without having to play the audio back. Nothing is hidden behind a play button.
When the Bot Replies With One
When a client writes to you by voice, the bot can reply with a voice note rather than text. It's a small thing that makes the conversation feel like a conversation — a client who spoke gets spoken back to.
Voice replies are not on by default. They require voice to be provisioned on your account and a plan that includes the feature. Your Remo account manager sets this up for you; there is nothing to configure yourself.
Both directions cost something. In your usage breakdown, incoming voice notes bill as Audio transcription and outgoing voice replies bill as AI audio reply, on top of the normal cost of the AI response itself.
AI Phone Calls
Some clients don't want to message at all. When a client explicitly asks to be called, the bot can place an actual outbound AI phone call — either straight away, or at a later time the client asks for.
What It Needs
- Voice calling provisioned on the project — this is arranged with your Remo account manager, not something you switch on yourself
- A phone number on the client's record — without a number there is nobody to call, and the bot falls back to messaging
How Repeat Requests Are Handled
Two safeguards keep a client from being pestered:
- Repeat requests are suppressed for a short window — if a client asks twice in quick succession, they are not called twice
- Only one scheduled call can be pending per client at a time — a new request for a later call replaces the intent rather than stacking a second call on top of it
Where It's Configured
Voice calling lives on Configuration → Advanced under Voice agent. Editing it is admin-only, and the initial setup is done together with your Remo account manager.
What the Bot Does With Other Media
Clients send all sorts of things into a WhatsApp thread. Here is what actually happens to each:
- Voice notes — transcribed and answered
- Photos, videos and stickers — saved to the conversation, but the bot does not reply to them
- Documents a client sends you — saved to the conversation, but the bot cannot read their contents
- Emoji reactions — the bot stays quiet for a simple 👍, and may respond to a reaction that reads as a question or a problem, such as a ❓
The pattern is consistent: everything is kept and visible to your team in the chat, but only voice notes are something the bot can act on. If a client sends a photo of a document or a screenshot with a question in it, someone on your team needs to pick that up.
Transcribing an incoming voice note and generating a spoken reply each consume credits on top of the AI response itself, so a voice-heavy conversation burns through a balance noticeably faster than a typed one. If you enable voice replies, keep an eye on your credit usage for the first few weeks and set your expectations from real numbers rather than your text-only history.