24.10. Response

Prev Next

The Response Assistant is the AI equivalent to the standard Speech component.

The Assistant will simply either read out your message word for word, or it will generate a message based upon your prompt.

Configuring a Response Assistant

After creating a Natterbox AI component from the Routing Policies, simply select the Response App.

Natterbox AI interface displaying various app options including Response and Knowledge.

Natterbox call response message indicating connection status for users.

  1. (Optional) Allow the AI to generate it’s own response

    1. If you need your message to be dynamic in nature, you can define a prompt using macros and allow the Assistant to create it’s own message on the fly

  2. Define the message the Assistant should relay

When to use a Response Assistant over Speech component

The two components are very similar, but currently the CAI Voices will only be available for the Response component, so to keep your user experience consistent it’s recommended to use a Response component if CAI is being used in your policies.

Initial Greeting

You may wish to use the Response Assistant to deliver the initial greeting to the customer. This allows you to control exactly what is said, rather than leaving it to the default assistant behavior.

Example Prompt:

Greet the customer warmly, introduce yourself, and offer to help. Add a small, delightful touch—such as a fun fact, a positive quote, or a light, situational joke—to create a welcoming and humanlike atmosphere.

Which may result in a response along the lines of

“Hi there! You’ve reached Emma, your AI assistant. I’ll do my best to help you today. Did you know honey never spoils? Archaeologists have found pots of it in ancient tombs, still perfectly edible! Now, how can I help you today?”

Once this initial greeting is delivered, the AI assistant will take over. It will be aware that the greeting has already occurred and won’t repeat it.

Dynamic Responses

Otherwise the main advantage the Response Assistant component has over the Speech App is the ability to dynamically inform your customers, such as summarising information from a macro.

For example you could fetch a Case from Salesforce and have your Assistant quickly summarise the contents back to the customer :

Summarize the case data in no more than two sentences: $(Custom_CaseData)

Complex Macros

If you wish to use a complex macro, such as a Salesforce query containing multiple results, then you’ll need to append ‘@JSON‘ for the Assistant to be able to fully access it, for example :

Summarize the case details to the customer. Only include the dateOpened and description for each case: $(SForce_Case.@JSON)