Flows provide a visual representation of a potential dialog in a user-friendly interface. Flows are composed of individual Nodes with different functionalities, ranging from generating output to collecting back-end information, and are executed as soon as an Input object comes in through an Endpoint.
How to create a new Flow
Want to learn how to build your first Flow? Have a look at Design a Flow and add a Message in our Help Center.
Flows are executed from top to bottom and use a logic-based system to determine which nodes should be executed. The execution path is shown through a highlighted green path.
Flow Nodes can be used to add interactivity to a flow. They can be added by clicking on the + icon in between two existing Nodes. There are many different types of Nodes, ranging from Question Nodes to Marketing Nodes.
Head over to Flow Nodes for more information on the different types of Nodes.
Flow execution typically starts at the top with the green Start Node . This is where the NLU takes place and the NLU results are subsequently published to the Input object. A flow is triggered with each new input and starts at the Start Node unless specified differently.
The Entry Point can change depending on the state of the conversation. It is indicated by a green triangle, next to a particular flow and shows where the next Input object (user input) would enter the flow. In case a Question Node has been triggered, the entry point will shift to that particular Node as can be seen in the following screenshot.
In some cases there are multiple Entry Points. This can be the case if an Optional Question was triggered or if a given Intent Entrypoint has been set explicitly.
In addition to the visual Chart, Flows also contain a tab named NLU. This section leads to the Natural Language Understanding that is relevant to this particular Flow.
NLU (Natural Language Understanding)
Please refer to NLU Overview to learn more.
Certain settings can be configured per Flow.
Information about the Flow Configuration Settings can be found here: Flow Settings.
Can be used as a starting point for the Flow's Context object. Head over to Context for more information.
Activate this setting to inherit the Intent model from fallback Locale. The Locale Settings tab is disabled for fallback Locale and enabled for alternative Locales. By default, the Inherit Intent model from fallback locale toggle is deactivated. Once activated, training indicator on the Intents page and error badge on NLU tab are hidden, also the Build Model button in Settings, Chart and NLU is disabled.
Switching to different Flows¶
With Cognigy.AI 4.10.0 switching between Flows with different settings and different Intents has ben improved as follows:
- When you are in the NLU tab of Flow A and you switch to Flow B you will still be in the NLU tab.
- When you are in the Settings tab of Flow A and you switch to Flow B you will still be in the Settings tab.