Phone & voice

A voice bot is an extra layer on top of your chatbot. You can reuse your existing chatbot, and we will handle speech-to-text (STT) and text-to-speech (TTS). Configuring voice using our platform is a matter of minutes and we've created some custom components to make the configuration even easier.

Activate your phone channel by going to the Channels tab and clicking the '+' button next to 'Phone'.

Architecture

Testing

If you want to test a first version of your voice bot before going into production, you can do so in two ways:

1 – Use the 'test your bot' feature by clicking on 'debugger' and then selecting voice. After you give your browser permission to use your microphone, you can now talk to your bot.

click on 'default' and select 'voice' to test your voice bot

2 – If you want to enact the real experience, you can call a test number. Before you can use a test number to call your bot, you have to add your number to the bot so we know which bot you're trying to talk to.

Adding your number: Click the blue 'Create Configuration' button to set up the phone configuration. The Chatlayer.ai test number will be shown at the top of the Phone configuration screen.

Production

When you're ready to move your bot into production, please get in touch. We will set up a specific phone number for your bot which can be used to call directly or forward calls to. A typical phone architecture looks like this:

Voice-specific Chatlayer.ai components

Silence detected event

If you want to trigger a certain flow when the user is silent for a set time, use the event which we describe here.

Voice message format type for input validation

You can save anything the user is saying as a variable by using the voice message format type in an input validation. Read more here.

Actions: forward call & close call

These actions are available when you configure the voice channel and will allow you to:

  • Forward a call to a number of your choosing, often used to handover a customer from the bot to a live agent

  • End the call by hanging up

SSML & external audio files

SSML is a markup language designed to give instructions to the text-to-speech engine. You can find more about SSML here. Chatlayer.ai supports the use of SSML in text messages which will be used by the TTS engine.

SSML can also be used to play external audio files in a text message. Alternatively, you can use the "media" template to achieve the same functionality.