OrderFood
might include sentences like Can you get me some food?
or I'd like to order a pizza
.
The NLU engine analyzes and scores these Intents to match user inputs.
Each Intent includes:
Working with Intents
- GUI
- API
You can create, edit, and manage Intents as well as download and upload them in NLU > Intents of the Flow editor.
Download and Upload Intents
Intents can be uploaded to and downloaded from Cognigy.AI as CSV or JSON files. Use these CSV or JSON files to import Intents, edit using your favorite spreadsheet or developer tools, and move them between Flows.CSV
CSV
Your CSV should be formatted as follows:
CSV format example:
- No header.
- Cognigy.AI automatically detects various CSV formats. If you encounter any issues, use UTF-8 encoding, comma (
,
) as the delimiter, double quotes ("
) as your quote character, and newline (\n
) as your line terminator. - You are required to fill in the following columns:
Name | Property type | Value |
---|---|---|
The name of the Intent used for identification and processing in Cognigy.AI. The Intent name is a string type value. | The property type corresponds to the following Intent components and settings:exampleSentence — a sample user input.defaultReply — the standard response when the Intent is triggered.confirmationSentence — a response that asks the user for confirmation.rules - the conditions to handle the Intent and guide the conversation based on the user input. | The value is a string type value and corresponds to an example sentence utterance, such as I want pizza . |
Keep regional settings consistent on your operating system (OS), as they affect the CSV delimiter. Mismatched settings can break the file and cause upload failure.
JSON
JSON
JSON format serves as a comprehensive specification for defining Intents within Cognigy.AI.The provided JSON example illustrates the structure and content of an Intent, including Intent name (
Pizza
), example sentences, rules, tags, disambiguation sentence, default reply, and other Intent-related settings.JSON format example:Add Example Sentences
When you create an Intent, add example sentences that help the NLU model recognize user queries.Build the NLU Model
Whenever you make changes to the Intent or its related settings, you need to rebuild the NLU model. If the NLU model needs rebuilding, the red indicator will be displayed:- next to the Flow name in the Flows list.
- next to the Build Model button in NLU > Intents of the Flow editor.
Single Flow
Single Flow
- In the left-side menu of the Project, go to Build > Flows.
- On the Flows page, select the Flow that you want to train.
- In the upper-right corner, select NLU.
- On the Intents tab, click Build Model or
> Quick Build.
All Flows
All Flows
- In the left-side menu of the Project, go to Build > Flows.
- On the Flows page, check if the Train all Flows button exists. If the Train all Flows button isn’t enabled for your environment, specify the
FEATURE_TRAIN_ALL_PROJECT_FLOWS: "true"
feature flag in thevalues.yaml
file for on-premises installations or contact Cognigy technical support. - Click Train all Flows.
Apply Additional Settings
Basic Additional Settings
Basic Additional Settings
Setting | Description |
---|---|
Inherit Example Sentences from Child Intents | Allows example sentences from child Intents to be inherited by the parent Intent. For more information, read Intent Hierarchy. |
Intent Mapping Priority | Defines the priority of the Intent when mapped against other Intents during the recognition process. For more information, read Intent Mapping Priority. |
Rules | Allows configuring rules to restrict Intent execution based on specific conditions. For more information, read Rules. |
Default Reply | Defines the fallback response if no other reply is configured or applicable. For more information, read Default Reply. |
Intent Entrypoint | Determines where the Flow execution starts when the Intent is triggered. You can select one of the following options: - Execute from START in this Flow — this option ensures that the Flow begins from the Start Node when the Intent is triggered. The option is selected by default. - Execute normally — the Flow execution continues from the Node where the Intent was recognized, without restarting the Flow. This option is useful when you want to maintain the conversation’s context and avoid restarting the Flow. - Execute from a Node in this Flow — the Flow execution starts from a specific Node when the Intent is triggered. This option provides flexibility to define custom Entrypoints within the Flow. |
Advanced Additional Settings
Advanced Additional Settings
Setting | Description |
---|---|
Tags | Allows assigning tags to the Intent for easier organization and filtering. |
Condition | Defines the conditions that must be met for the Intent to be executed. For more information, read Conditions. |
Confirmation Sentence | Specifies the sentence used to confirm the Intent with the end user. For more information, read Confirmations. |
Disambiguation Sentence | Provides a sentence to clarify the Intent if there is ambiguity during recognition. The disambiguation sentence will be exposed in the Input object: - input.nlu.intentMapperResults.finalIntentDisambiguationSentence — the disambiguation sentence for the best-matching Intent that was selected.- input.nlu.intentMapperResults.scores[0].disambiguationSentence — the disambiguation sentence for a specific Intent that was matched. |
Analytics Step | Marks the Intent for analytics tracking, allowing it to be included in analytical reports in Insights and OData. |