GoSmooth

Feature · Voice notes

An AI sales agent that understands voice notes

Book a demo

30 minutes · on your own catalog · no commitment

  • On WhatsApp, Instagram and Messenger
  • Answered in writing, in the buyer’s language
  • Photos and videos understood too

Understood in a voice note

A voice note, 9 seconds0:09

What the agent heard“Hi, do you have the floral blouse in M? How much to Lekki?”

  • English
  • French
  • Arabic
  • Darija, mixed with French
  • Nigerian Pidgin
  • Hausa
  • Yoruba
  • Igbo

Up to 25 MB a voice note. The answer comes back in writing, with the product, the price and the delivery.

How the agent answers a voice note

From “how much to Lekki?” to a confirmed order, without the buyer typing

Three moments

Where a voice note loses the sale

Not in what the buyer said. In how long it waits to be heard.

One switch flips the three moments.

  • A voice note at 23:00

    You listen in the morning, between other chats, and type the answer.

    Heard at 09:30

    Transcribed and answered in writing, at the hour it arrives.

    Answered at 23:01
  • A voice note you can’t make out

    You guess what they said, or ask them to type it out.

    A guess

    If it can’t be transcribed, the agent asks the buyer to say it again.

    Asked again, not guessed
  • A photo instead of a product name

    You scroll your own posts to work out what they mean.

    Found in your grid, later

    The photo reaches the agent as a picture, beside the buyer’s words.

    Seen with the message

In the chat

What the agent does with what buyers send

Voice notes, photos and videos each reach the agent in their own way. The answer is always written.

  • On all three channels

    Voice notes on WhatsApp, Instagram and Messenger are transcribed the same way.

    WhatsApp · Instagram · Messenger
  • In the buyer’s language

    Transcribed in the language the buyer speaks, and answered in writing in that language.

    English in, English out
  • Up to 25 MB

    Each voice note can be up to 25 MB. A very long one is read up to 4,000 characters of text.

    25 MB a voice note
  • Photos, as pictures

    A photo reaches the agent as a picture, beside what the buyer wrote.

    Photo + “is it this colour?”
  • Videos, described

    A video reaches the agent as a description and up to three frames.

    A description + 3 frames
  • Asks again

    When a voice note can’t be transcribed, the agent asks the buyer to say it again.

    “Could you say it again?”
  • Then the order

    From a voice note to the recap: the product, the price, the delivery for the city, Confirm.

    ORD-2026-3714 · ₦21,000
  • Written replies only

    It answers in writing, not with voice notes. A PDF the buyer sends is kept, not read.

    Voice replies: not yet

FAQ

Voice note questions sellers ask

Short answers first. The full answer is one tap away.

Yes. Voice notes on WhatsApp, Instagram and Messenger are transcribed the same way, up to 25 MB each.

If it can’t be transcribed, the agent asks the buyer to say it again. It doesn’t answer a voice note it couldn’t hear.

Not yet. It answers in writing, in the buyer’s language, so the price and the delivery stay on screen.

A voice note can be up to 25 MB. A very long one is read up to its first 4,000 characters of text.

Photos reach the agent as pictures, and videos as a description with up to three frames. PDFs are kept, but the agent can’t read them.