What does a Flow consist of?
Intent Trigger, transitions, output and nodes
A Flow is a graphical representation of a dialogue between a conversational AI application and an end-user. It consists of several nodes linked together with arrows (transitions) showing the direction and the logic of the Flow. A Flow is activated by one or more Flow triggers. The Flow structure contains Flow nodes of various types that are linked by transitions. Output nodes contain the answers that the conversational AI application will provide.
The image above presents an example of a Flow made up of the following:
- Intent Trigger: a Flow can have one or more Triggers, and each Trigger can have one or more Match Requirements as well as Data Actions.
- Output node: this node contains the conversational AI application's answer(s).
- "Get Input" transition: these transitions force the Flow processing to halt until the user provides new input. If the user input matches the condition set on the transition (of if the transition is unconditional), the Flow processing will resume along this transition.
- Script node: the script nodes allow to add programming code to the Flow. This node for example calculates the number of days (from today) until the visitor's next birthday.
- "Continue" transition: these types of transitions are used to link one node to another when no new user input is required.
Other nodes that might appear in the Flow graph:
- Flow links: the Flow links allow to link Flows together, thus making it possible to combine processes that are implemented in different Flows. To add a Flow link, click the corresponding icon in the top ribbon of the Flow or right-click the Flow and select the corresponding option in the context menu. Read more in the Linking Flows section.
- Junctions: the junctions allow to split the Flow of the dialogue; they are often used immediately after the Flow trigger when the initial reply of the conversational AI application depends on some string(s) in the user input.