Interactions: Events
A client request sent to our API triggers an event. No user input is needed; the activation of browser/application code alone is sufficient to be identified as an event. An event doesn’t happen by itself. It needs to be pre-defined in Conversational AI Cloud.
(e.g. loading page x is an event, but loading pages y and z are not) and is only active when someone (through our JavaScript or through the front end code) ensures that the event call is actually happening.
Calling the event through Conversational AI Cloud itself in response to other interactions with Conversational AI Cloud can be done, but these are not the only events possible. To call an event externally, Conversational AI Cloud's API has an event endpoint.
Naming of Events
Event names can only contain 256 characters and only contain alphanumeric characters, underscores (_) and/or hyphens (-) and no spaces.
The event name should indicate its purpose. For example:
onLoad: events which load on a page (before any customer interactions)
Exception: events that activate before an interaction has been fully processed
Escalation: events that activate because of the result of other interactions in Conversational AI Cloud
Events Response Output
We use the word ‘response’ for any output of Conversational AI Cloud. There are different kinds of responses (outputs) logged:
A technical response code, accompanied with a description or not.
An answer i.e. text output, possibly containing URLs, images, videos etc.
A page push; a URL that the client will refer to automatically, without any action from the end-user.
A Video, i.e. a URL referencing the video.
A list of FAQs, i.e. a collection of questions
A combination of these can also occur as a more complex response. Please view our developer documentation for more information concerning any possibilities and limitations.