Dialogue consists of intent and entity.

1. Go to Dialogue Setting Page

1.1 First Approach

• In my robot page, click on Settings of the chatbot card to enter the skills configuration page

• Toggle to the right the small switch on the dialogue card to make sure the skill is active

• Click Editor on the dialogue card to enter the NLU interface

1.2 Second Approach

• In My Skills page, click Editor on skill cards to directly enter the NLU interface

1.3 Difference

• Both ways can operate on NLU, but the difference is that

  • The first approach is to operate the skills at the chatbot level. After the operation is completed, the chatbot is trained and then tested in dialogue with the chatbot

  • The second approach is to operate the skills at the question and answer pair level. After completing the operation and training, the dialogue test is conducted based on the skills

• They are essentially indistinguishable. Recommend the use of the first approach in the early stages, as the documentation is based on the first approach

2. Intent

• Intent is used to deal with more complex intention, question templates, and answers
• Users can edit customized intent conversations in the Intent page
• The chatbot will be trained to match the user-defined question template with the corresponding template answer

2.1 Add Intent

• Click Add Intent and enter name and description to add a new intent dialog template

• Click Save

2.2 Edit Intent

• In the intent list, click the intent name to enter the corresponding intent

• Edit area includes template, answer, and intent slots

2.2.1 Edit Template

• Templates are used to simulate conversations that users might enter

Users can write a dialog template for the chatbot through dialogue intent. Dialogue templates include question templates and answer templates Use “+ enter template” to enter a new template, and press Enter to add a new template Use “+ enter template answer” to enter a new template answer, and press Enter to add a new template Click the question/answer in the template again to edit the question/answer (press Enter to save)

2.2.2 Edit Template Answer

• The template answer is the chatbot's reply to the template question

2.2.3 Add Intent Slots

• Intent slots can be used to annotate words in template question and template answer statements

• Users can enter a new slot name in the slot table, the entity on which it is based (a required option), and click add flag (or press Enter after entering the name) to add a new slot
• Click the slot name again (such as 'place') to edit the slot name
• In Select a slot type the entity can be selected or reselected based on the slot

2.2.4 Mark Slot

• The way to mark the slot is to select a segment of the sentence answered by the template/template with the mouse. The slot option will pop up immediately. Click to select a slot

• Multiple slots can be added per template
• A slot can be added for both template questions and template answers
• After the slot is successfully added, the original template content will be placed into the corresponding slot color block

2.2.5 Delete Slot

• Users can delete the slot

• The deleted slots will automatically lose their validity in the template

2.3 Training and Testing

• Skills must be trained after editing. Click on Train in the upper right corner

• Then test it in the dialog window

2.4 Fulfillment

• Dialogue can be responded to template questions by a customized fulfillment
• After Fulfillment is enabled, the user can enter the URL of the fulfillment API, and the template answer function and script function will be disabled

• With Code Editor enabled, users can edit scripts to support Groovy and Python, while template answer and fulfillment are disabled

• If any of the above functions are enabled, the template answer function is disabled

2.5 Import / Export

• Users can import dialogue data in bulk, and the imported data will overwrite the dialogue content added by users in the platform
• Click on View sample file to view the sample file for uploading

• Click Export, a confirmation box will appear. Enter the file name and click export to download the current intent conversation skills into a JSON file

3. Entity

• Sometimes users use synonyms to express the same query, but the system cannot recognize them, for example

• Entity can easily solve this problem
• Entity is used to define and describe the types and contents of slot words
• When the users set the slot, in addition to filling in the slot name, they also need to select the entity on which the slot is based on

• Click Add Entity to add customized entity

• Users can add a series of customized entities to the entity page. They can also edit and delete entities

• Now go back to the Intent screen and use it. The user-defined entity will appear under the Userdef list in the slot type when the user adds a slot

• Label the slots

• Remember to click on training after modification. Then test it in the dialog window