Chatbot Response Types
The Medium Impacts the Message: Communicating using text, images, buttons and even pauses.
There is a certain allure for the purest-minded when in comes to Natural Language. And that is to keep the user input to just that; natural language. Apart from the beauty of the simplicity of it, there are also some practical considerations for this.
The one is advantageous is that basic text is medium agnostic. Hence a text only interface can live on a medium like SMS/Text or any other medium for that matter. No special handsets or applications are required and the interface can truly be ubiquitous. The user does not need to have a smartphone, feature phone etc.
Especially in the third world context where basic or feature phones are prevalent and no or significant proliferation of smartphones / feature phones exist.
However, when your conversational interface or chatbot exists on a medium which is feature rich, like Facebook Messenger, WhatsApp, Twitter and the like, you will most definitely be tempted to introduce richer and varied responses. Think of examples where you want to display a menu, or a clothing site.
Here is an example of an IBM Watson Assistant implementation of a richer more interactive user interface. This demo is recorded from the test interface within the cloud development environment.
A dialog node can respond to users with a message which include text, images, or interactive elements such as clickable options. If you are building your own client application, you must implement the correct display of all response types returned by your dialog.
It is the responsibility of your client application to handle all response types appropriately. In this case, your application would need to display the specified text and image to the user. This is where the medium impact the message and the way dialog data is represented is dictated by the payload definitions as defined by your medium of choice. The medium of choice being, for example, Messenger, WhatsApp, Twitter, web chat etc.
Within the IBM Watson Assistant environment there are three building blocks or elements which constitute the application. These are intents, entities and lastly the dialog. The dialog can bee seen as the state machine. Within the dialog there are nodes, and each node can have a defined response type.
Here the responses available are Text, Option, Pause and Image. The text response type is used for ordinary text responses. The image response type instructs the client application to display an image, optionally accompanied by a title and description.
The pause response type instructs the application to wait for a specified interval before displaying the next response; defined in milliseconds.
A pause response is typically sent in combination with other responses. Your application should pause for the interval specified by the time property (in milliseconds) before displaying the next response in the array. The optional typing property requests that the client application show a "user is typing" indicator, if supported, in order to simulate a human agent.
The option response type instructs the client application to display a user interface control that enables the user to select from a list of options, and then send input back to the assistant based on the selected option
Read more here…
IBM Watson Assistant Define Responses: