Defining responses using the JSON editor

Fields

Example

This example displays an image with a title and descriptive text.

{
  "output": {
    "generic":[
      {
        "response_type": "image",
        "source": "https://example.com/image.jpg",
        "title": "Example image",
        "description": "An example image returned as part of a multimedia response."
      }
    ]
  }
}

Fields

Example

This example displays an video with a title and descriptive text.

{
  "output": {
    "generic":[
      {
        "response_type": "video",
        "source": "https://example.com/videos/example-video.mp4",
        "title": "Example video",
        "description": "An example video returned as part of a multimedia response."
      }
    ]
  }
}

Fields

Example

This example plays an audio clip with a title and descriptive text.

{
  "output": {
    "generic":[
      {
        "response_type": "audio",
        "source": "https://example.com/audio/example-file.mp3",
        "title": "Example audio file",
        "description": "An example audio clip returned as part of a multimedia response."
      }
    ]
  }
}

Fields

Example

This example embeds an iframe with a title and description.

{
  "output": {
    "generic":[
      {
        "response_type": "iframe",
        "source": "https://example.com/embeddable/example",
        "title": "Example iframe",
        "description": "An example of embeddable content returned as an iframe response."
      }
    ]
  }
}

Fields

Example

This example displays two options:

• Option 1 (labeled Buy something) sends a simple string message (Place order), which is sent to the workspace as the input text.
• Option 2 (labeled Exit) sends a complex message that includes both input text and an array of intents.

{
  "output": {
    "generic":[
      {
        "response_type": "option",
        "title": "Choose from the following options:",
        "preference": "button",
        "options": [
          {
            "label": "Buy something",
            "value": {
              "input": {
                "text": "Place order"
              }
            }
          },
          {
            "label": "Exit",
            "value": {
              "input": {
                "text": "Exit"
              },
              "intents": [
                {
                  "intent": "exit_app",
                  "confidence": 1.0
                }
              ]
            }
          }
        ]
      }
    ]
  }
}

Fields

Example

This examples sends the "user is typing" event while pausing for 5 seconds.

{
  "output": {
    "generic":[
      {
        "response_type": "pause",
        "time": 5000,
        "typing": true
      }
    ]
  }
}

Fields

Example

This examples displays a greeting message to the user.

{
  "output": {
    "generic":[
      {
        "response_type": "text",
        "values": [
          { "text": "Hello." },
          { "text": "Hi there." }
        ],
        "selection_policy": "random"
      }
    ]
  }
}  

Fields

You can author a search_skill response_type that calls out to your search skill, and then return a search response_type that you can render in a different format. Go to Message > Response > output > MessageOutput > generic > RuntimeResponseTypeSearch in Watson Assistant v2 API to view how response_type search renders when search_skill response is processed..

Example

This example uses the user input text to send a natural-language query to the search skill.

{
  "output": {
    "generic": [
      {
        "response_type": "search_skill",
        "query": "",
        "query_type": "natural_language"
      }
    ]
  }
}

Fields

Example

This example requests a transfer to a live agent and specifies messages to be displayed both to the user and to the agent at the time of transfer.

{
  "output": {
    "generic": [
      {
        "response_type": "connect_to_agent",
        "message_to_human_agent": "User asked to speak to an agent.",
        "agent_available": {
          "message": "Please wait while I connect you to an agent."
        },
        "agent_unavailable": {
          "message": "I'm sorry, but no agents are online at the moment. Please try again later."
        }
      }
    ]
  }
}

Fields

Example

This example requests a transfer from Slack to web chat. In addition to the channel_transfer response, the output also includes a text response to be displayed by the web chat integration after the transfer. The use of the channels array ensures that the channel_transfer response is handled only by the Slack integration (before the transfer), and the connect_to_agent response only by the web chat integration (after the transfer). For more information about using channels to target specific integrations, see Targeting specific integrations.

{
  "output": {
    "generic": [
      {
        "response_type": "channel_transfer",
        "channels": [
          {
            "channel": "whatsapp"
          }
        ],
        "message_to_user": "Click the link to connect with an agent using our website.",
        "transfer_info": {
          "target": {
            "chat": {
              "url": "https://example.com/webchat"
            }
          }
        }
      },
      {
        "response_type": "connect_to_agent",
        "channels": [
          {
            "channel": "chat"
          }
        ],
        "message_to_human_agent": "User asked to speak to an agent.",
        "agent_available": {
          "message": "Please wait while I connect you to an agent."
        },
        "agent_unavailable": {
          "message": "I'm sorry, but no agents are online at the moment. Please try again later."
        },
        "transfer_info": {
          "target": {
            "zendesk": {
              "department": "Payments department"
            }
          }
        }
      }
    ]
  }
}

Fields

The command_info.type field can specify any of the following supported commands:

• collect: Collects DTMF keypad input.
• disable_barge_in: Disables DTMF barge-in so that playback from the phone integration is not interrupted when the customer presses a key.
• enable_barge_in: Enables DTMF barge-in so that the customer can interrupt playback from the phone integration by pressing a key.
• send: Sends DTMF signals.

For detailed information about how to use each of these commands, see Handling phone interactions.

Example

This example shows the dtmf response type with the collect command, used to collect DTMF input. For more information, including examples of other DTMF commands, see Handling phone interactions.

{
  "output": {
    "generic": [
      {
        "response_type": "dtmf",
        "command_info": {
          "type": "collect",
          "parameters": {
            "termination_key": "#",
            "count": 16,
            "ignore_speech": true
          }
        },
        "channels": [
          {
            "channel": "voice_telephony"
          }
        ]
      }
    ]
  }
}

Fields

Currently, the following activities for the phone integration can be stopped:

• speech_to_text_recognition: Stops recognizing speech. All streaming audio to the Speech to Text service is stopped.
• dtmf_collection: Stops processing of inbound DTMF signals.

Example

This example uses the stop_activities response type to stop recognizing speech. Because this command is specific to the phone integration, the channels property specifies voice_telephony only.

{
  "output": {
    "generic": [
      {
        "response_type": "stop_activities",
        "activities": [
          {
            "type": "speech_to_text_recognition"
          }
        ],
        "channels":[
          {
            "channel":"voice_telephony"
          }
        ]
      }
    ]
  }
}

Fields

Currently, the following activities for the phone integration can be started:

• speech_to_text_recognition: Starts recognizing speech. Streaming audio to the Speech to Text service is resumed.
• dtmf_collection: Starts processing of inbound DTMF signals.

Example

This example uses the start_activities response type to restart recognizing speech. Because this command is specific to the phone integration, the channels property specifies voice_telephony only.

{
  "output": {
    "generic": [
      {
        "response_type": "start_activities",
        "activities": [
          {
            "type": "speech_to_text_recognition"
          }
        ],
        "channels":[
          {
            "channel": "voice_telephony"
          }
        ]
      }
    ]
  }
}

Fields

The command_info.type field can specify any of the following supported commands:

• configure: Dynamically updates the Text to Speech configuration. Configuration changes can be applied only to the next conversation turn, or for the rest of the session.
• disable_barge_in: Disables speech barge-in so that playback from the phone integration is not interrupted when the customer speaks.
• enable_barge_in: Enables speech barge-in so that the customer can interrupt playback from the phone integration by speaking.

For detailed information about how to use each of these commands, see Applying advanced settings to the Text to Speech service.

Example

This example uses the text_to_speech response type with the configure command to change the voice used by the Text to Speech service.

{
  "output": {
    "generic": [
      {
        "response_type": "text_to_speech",
        "command_info": {
          "type": "configure",
          "parameters" : {
            "synthesize": {
              "voice": "en-US_LisaVoice"
            }          
          }
        },
        "channels":[
          {
            "channel": "voice_telephony"
          }
        ]
      }
    ]
  }
}

Fields

The command_info.type field can specify any of the following supported commands:

• configure: Dynamically updates the Speech to Text configuration. Configuration changes can be applied only to the next conversation turn, or for the rest of the session.

For detailed information about how to this command, see Applying advanced settings to the Speech to Text service.

Example

This example uses the speech_to_text response type with the configure command to change the language model used by the Text to Speech service to Spanish, and to enable smart formatting.

{
  "output": {
    "generic": [
      {
        "response_type": "speech_to_text",
        "command_info": {
          "type": "configure",
          "parameters": {
            "narrowband_recognize": {
              "model": "es-ES_NarrowbandModel",
              "smart_formatting": true
            }
          }
        },
        "channels":[
          {
            "channel": "voice_telephony"
          }
        ]
      }
    ]
  }
}

Fields

For the phone integration, you can use the channel_options object to include custom headers with the SIP BYE request that is generated. For more information, see End the call.

Example

This example uses the end_session response type to end a conversation.

{
  "output": {
    "generic": [
      {
        "response_type": "end_session"
      }
    ]
  }
}

Fields

Example

This examples shows a generic example of a user-defined response. The user_defined object can contain any valid JSON data.

{
  "output": {
    "generic":[
      {
        "response_type": "user_defined",
        "user_defined": {
          "field_1": "String value",
          "array_1": [
            1,
            2
          ],
          "object_1": {
            "property_1": "Another string value"
          }
        }
      }
    ]
  }
}