Intents and Entities Creation

Intents and Entities Creation

Step 1. To manage your AI Model capabilities, go to the application of AI Model Configurator.

Step 2. Select your AI Model from the dashboard.

Step 3. To start creating model data, go to the tab of Conversational AI and select Intents to create a new intent. To create a new intent, press the Add Intent button.

Step 4. Fill the intent name, adjust strictness level and press the Add button.

Step 5. Choose the newly created intent and add expressions to it.

Step 6. To create a new entity in the Conversational AI tab, choose Entities and press the Add Entity button:

Step 7. Enter the entity name and select its type (Free, Restricted or Regex), adjust the strictness level.

Step 8. Add Values to your Entity.

Step 9. Supplement the entities with enrichments (Optional).

To add Enrichments in the Enrichments tab, press the Add Enrichments button

To configure your enrichment, go to Edit mode, press the Add New Pair button, add Key and Value, save the changes.

Here we created entity: contact.

Example: Every time our AI Model will recognize entity contact in the user’s message, it will also add to the memory those key-value pairs: (source: website, and priority: standard) .

Following our example, we want source of the contact to be set as “website”, and we set “priority” as “standard”.

We can also set custom enrichments for only some entities.

Press the Add New Group button, enter the Group name (1). You will see the enrichment key-value pairs you have set in previous step (2).

For the entity values: call and phone(3) let’s change the “priority” to “high”.

Now every time a user will mention contact by phone or call, the priority will be set to high in memory.

Step 10. Tag Entities in Intent expressions.

You have a possibility to Tag entities in Intents. To do so, go to the Conversational AI/Intents tab, select your intent.

Press the Tag icon near the expression where you’d like to add Tags.

Select Entity from the dropdown list. Then choose the word in the expression to tag it as a selected entity.
Save your entries.

To delete a tag, just press on it one more time, so the colorful highlight disappears. Save your changes.

To add changes or delete the tag after it was saved, click on the appropriate entity tag in the list of Intents.

Conversational AI Settings

Step 1. Edit the Settings.

Open the Settings tab (1) in the Conversational AI section on the left. Click on the information icon (2) to understand the options. Click on the Edit (3) button to change them.

Intents:

Matching Strictness – strictness ranges from 1 to 100, where 1 is not strict , and 100 is maximum strictness. Model set to 90-100 strictness will base on the training data for intents and entities exclusively, and will not recognize words that are a bit modified – like typos, plural forms, verbs in different tenses etc. Lower strictness will allow for more elasticity.
Recommended value: No less than 90 for optimal accuracy.

Restricted Entities:

Matching Strictness: Define the precision level for recognizing restricted entities. Definition is the same as for Intents.
Recommended value: No less than 90 for effective filtering.

Gold Entities:

Choose Recognition Type:

Embedded: Uses LLM library-based recognition logic. Suitable for simpler, localized recognition needs.
Generative AI-Based: Uses advanced Generative AI for greater flexibility and robust recognition. Recommended for complex or dynamic scenarios.

Configure Settings Based on Recognition Type

For Generative AI-Based Recognition: Select a Recognition Destination from the available Generative AI Destinations.
For Embedded Recognition: Set the Matching Strictness: Recommended value: 90 or higher for accurate recognition. Higher values ensure precise matches; lower values broaden interpretation.

Language Controls: