Links

Voice

Follow these step-by-step instructions on how to set up Sinch voice channel and voice settings

Set up voice channel

  • Go to the 'Channels' tab on the left-side menu and click on the 'Add channel' button, search for Voice, select it and click 'Next':
You will see the message below. Click on 'Create':
The channel is created and you can now click on 'Setup Sinch Voice':
You will be requested the Sinch Key and App secret.
You will need to copy Key and Secret (in the card called 'Your summary') from your voice App on Sinch Dashboard...
...and paste it in Chatlayer:
When the previous steps are completed, you'll be able to see the same number from your Sinch Dashboard Voice App in your numbers for the Sinch Voice channel on Chatlayer. The Callback URL from Sinch Dashboard will be automatically filled in for you.

Choosing a voice for your bot

  • Click on 'Settings' in the left menu and select 'voice'
  • Under 'Configure speech to text', select the language of your bot. You can also select the correct dialect if you'd like by clicking on the box 'Default dialect'
  • Under 'Voices', select the voice you prefer
  • If you'd like to hear a sample of that voice, enter an example sentence in the box 'Text to try out' and click on the speaker that appears next to the voice to hear it read aloud

Customizing your bot's voice with SSML

You can customize your bot's responses by using Speech Synthesis Markup Language (SSML) and enhance the conversational experience. A few examples of how you can tweak your bot's responses with SSML:
  • Change certain pronunciation​s (especially handy when you have a brand name that needs to be pronounced a certain way)
  • Change the speed​ of your bot's voice (for example, when giving a number that the user needs to write down, or spelling something)
  • Play audio file​s
  • Add pauses (natural breaks between blocks of information, just like a real human would do!)
  • Pronounce words as characters
  • Pronounce dates
There's much more cool stuff to do with SSML! See more here: https://cloud.google.com/text-to-speech/docs/ssml​
To enable SSML, you need to put <speak> … </speak> around the entire bot text.
Using a Readspeaker (Nuance) voice? Then you need a different syntax. For example, when building a Flemish bot, this would be the SSML syntax: <speak version="1.1" xml:lang="nl-BE">Goedemorgen</speak>
Read more about Readspeaker's SSML guidelines here: https://docs.readspeaker.com/assets/markup/ReadSpeaker_SSML_Documentation_v1-3_01.pdf​

Forwarding and closing a call​

  • Add an 'Action' dialog and select 'Forward Call'
  • In the 'To' box, enter the phone number you'd wish to redirect the call to
The phone number should contain a country code but no leading zeros, only the '+' sign. For example: 0800 55 800 becomes +32 800 55 80
  • After setting up the phone number you wish to forward your call to, the bot will automatically hang up upon transferring. To reach the bot again, hang up the call and call a second time.

DTMF

Dual-tone multi-frequency, or clicking the numbers on the telephone dialpad to be redirected to a specific queue or option, is enabled on Chatlayer. In the use case below, we'll make use of speech recognition or DTMF to redirect the user to the correct queue.
These are the caller options:
  • Say Sales or press 1 to be transferred to the Sales department
  • Say Marketing or press 2 to be transferred to the Marketing department
  • Say Support or press 3 to be transferred to the Customer Support department
The user input is saved in the variable choice, which will be matched to the Go To options and redirected to their respective bot flow (Actions that will forward the calls to each queue):
Using speech recognition:
Using the telephone dialpad:

Call recording

If you have a voice-bot, chances are you need to record part of your calls for legal or administrative reasons, so here’s how you can do it:

A – If you only want to record a specific part of the flow, or from a point of it to the end (it may include the conversations with live agents when Offloading to Sinch Contact Pro, if that is your offloading provider).

By default, your Sinch Voice channel settings for recording will be Off:
Channels > Sinch Voice channel > Settings > Toggle OFF
While editing your bot, when you reach the point in the flow where you want to start the recording, add a start recording action:
The widget will show up in the Action bot dialog:
The following step is optional and dependent on your tests
If you realize that the last bot or user utterance is still showing up in the recording, we suggest you add a delay, before the start recording action, that is long enough so the last message before the action is not recorded. Make sure to add the delay before the start recording action:
If you wish to stop recording from one point in the bot:
If you want to prevent your recording from including the conversation between the user and live agent from Sinch Contact Pro after offloading, add a stop recording action prior to transferring:
If you wish to maintain the recording until the end of the live agent conversation, you don’t need to add a Stop recording action.
Please bear in mind that even if you add more than one start/stop recording actions in the same bot, only the last one will be available in the recording widget.

B - You want to record the call from the moment of first connection:

You can achieve that by accessing: Channels > Sinch Voice channel > Settings > Toggle ON
With this toggle, the whole conversation will be recorded and you can listen to it from the History of Conversations

Sending call attached data to Sinch Contact Pro

It is possible to send call attached data (CAD) when transferring a call to Sinch Contact Pro. That will help ensure the compliance with data privacy laws and allow quality assurance and audit of the calls on the Contact Pro platform.
From the moment on the bot flow where the consent to being recorded or not has been identified, you can redirect the user to an empty bot message with the variables RecordingEnabled and RecordingConsentGiven with the values 'yes' or 'no' depending on the will of the caller.
From an active call on the Communications panel from Sinch Contact Pro, you can access the Conversation Details widget, on the right side menu:
You can also check the Chatlayer Transcript widget to read the bot transcripts and listen to the recorded part of the bot, if there's any:
Here you can see a screenshot of the widget with an audio clip from the chatbot in an ongoing conversation and in the Wrap up page from Sinch Contact Pro:
In both of these cases:
  • call recording is not enabled for a part or the etire bot conversation
  • when the whole call (bot + live agent conversation) is being recorded
you won't be able to listen to the recording. The widget will appear empty like shown in this image:

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.
​