Usecase Design Console (UCD Console)

The use case design console (also known as the UCD console) is the primary area where conversation design is done for a particular skill or tenant. Users can create new skills, intents, entities and finally the dialog nodes to handle user utterances and customize conversational experience.

This console helps to configure the usecases to interact with the customers in a useful way by addressing their queries and providing suitable responses.

1. Go to Admin Dashboard Console under BigFix AEX Cognitive Consoles and click on Usecase Design Console.
2. The Usecase Design Console appears.

   

3. Select SKILLS tab. Two buttons are available under this tab:
   • Create Skill
   • Import and Create Skill
4. Click on Create Skill button under SKILLS tab to create a new skill.
5. The Create Skill form appears:

   

6. The following fields need to be populated:
   • Skill Name: Enter the name of the new skill.
   • Description: Provide brief description about the skill.
   • Language: Select the preferred Language from the drop-down.
7. Click on Add button.
8. The new skill will be created and will be visible under the SKILLS tab.

   

9. Click on OPEN on the newly added skill.
10. The following screen appears with new tabs under the newly added skill:

    

11. Click on IMPORT AND CREATE SKILL button under the SKILLS tab to create a new skill with the existing intents, entities, and dialog nodes.
12. It prompts you to Import the JSON file.

    

13. Click on BROWSE button and select a JSON file that you want to upload.
14. Click on Import File.
15. Once the file is imported a success message appears and the imported intents, entities and dialog nodes become visible under INTENT, ENTITY and DIALOG NODES tabs respectively.
16. Click on OPEN button present on each skill to open an existing skill.
17. To view the entities, intents and dialog nodes associated with the skill click on ENTITY, INTENT and DIALOG NODES respectively.

    

    

18. The Kebab menu (the three dots menu) present on each skill allows you to open a menu with additional options. The following options are available:
    • Export: Allows to download the JSON file associated with a particular skill.
    • Import: Allows to import the existing intents, entities, and dialog nodes for a particular skill. Follow the steps explained above ( Step 11 to Step 13 ) to perform the import action.

    

19. MULTISKILL on Usecase design enables switching between the multiple skills. The console facilitates the user to be able to add skills which he chooses to switch between and choose a default skill for the conversation. The Multi Skill Orchestration includes the default method, waterfall method, broadcast method and the conductor-follower method. In default method there are options to select the skill while in other methods the transition between skills would be automated.
20. Below is an explanation of how each method of skill orchestration differs and is similar to the others:
21. • Default Model:
      • This is the already existing (and till now the only) method in multi skill functionality. Here we get options to select multiple skills in the use case design console and associate a button name to it and select one default skill among the selected skill.
      • In the chat window we get the options to choose from the skill that we saved in UCD multi skill and further communication continues with the selected skill.
    • Waterfall Model:
      • It is a priority-based method i.e., the skill with highest priority would be checked first. If it provides the correct answer (doesn’t reaches to the Anything else node) then BigFix AEX does not go to the subsequent skills.
      • If execution reaches to the last skill and still reaches Anything else of the last skill, the message associated with that node would be returned. On the Use Case Design Console, the user selects the skills and their priorities.
    • Broadcast Model:
      • This method is based on the Highest confidence score of skill i.e., BigFix AEX checks all the skill at once and provides response from the skill that has the highest confidence score.
      • In this method, we select a default skill, and that skill provides us with the welcome message, also in cases where we don’t get appropriate response, we will have Anything else node message from this skill.
      • On Use Case Design Console, the user selects the skills and chooses one default skill.
    • Conductor-Follower Model:
      • In the Conductor-Follower method, there are 2 types of skill selection,
      • conductor skill: In this skill, the user adds 1 skill.
      • follower skill: In this skill, the user has multiple categories to add.x
      • The conductor skill acts as a default skill and is responsible for the transition of skill i.e., based on the user’s query, the conductor skill selects a follower skill for further communication.
      • The conductor skill, based on specific conditions, sends the id of the selected follower skill as part of intent and this id in turn is used to call the skill for further communication.

Multiskill Configuration – Default Model

Perform the following steps for MULTISKILL configuration using the Default Model:

1. Go to the BigFix AEX Tenant and open use case design Console.

   

2. After Opening the use cases design Console, you can find the Multi skill button. Click Multiskill.

   

   Note:

   In the Usecase Design console the MULTISKILL feature is visible only if the instance is enabled with the multi skill configuration.

3. The Multi Skill Configuration page appears. The page lists all the available models. Select the model and the console displays the corresponding options based on the selection.
4. Select the Default Model and click ‘Continue’.

   

   Note:

   If user selects the None option in the Multi skill Configuration page, by default, it takes the default model for the skills to be mapped for the tenant.

   

5. In Default Model, populate the Skills field by using the Choose a Skill drop-down. Multiple skills can be added in the Skills field. Enter the Button Name for the skill. Select the Default Skill by clicking Choose Your Option drop-down.

   

   Note:

   The Button Name must be unique in the default model.

6. Click Save to configure the multi skill.

   

7. The following confirmation message appears:

   

8. Click Ok to save the configuration. The following success message appears:

   

9. To delete a skill in the default model, click on the delete icon corresponding to the skill to be deleted.

   

10. The following confirmation message appears. Click Ok to confirm the action.

    

Multiskill Configuration – Waterfall Model

• Perform the following steps for MULTISKILL configuration using the Waterfall Model:

1. In the Multi Skill Configuration page, select Waterfall Model and click Continue.

   

2. In the Waterfall Multi Skill Configuration, Select Skills by using the Choose a Skill drop-down.

   

3. In the waterfall model, user has the options to move the skill up and down. To move a skill up, click the Move up arrow and to move a skill down, click the Move down arrow .

   

   

4. Click Save to configure the multi skill.

   

5. The following confirmation message appears:

   

6. Click Ok to save the configuration. The following success message appears:ca

   

7. To delete a skill in the waterfall model, click on the delete icon corresponding to the skill to be deleted.

   

8. The following confirmation message appears. Click Ok to confirm the action.

   

Multiskill Configuration – Broadcast Model

• Perform the following steps for MULTISKILL configuration using the Broadcast Model:

1. In the Multi Skill Configuration page, select Broadcast Model and click Continue.

   

2. In Broadcast Multi Skill Configuration, select Skills using the Choose a Skill drop-down. Select the Default Skill by clicking Choose your option drop-down.

   

3. Click Save to configure the multi skill.

   

4. The following confirmation message appears:

   

5. Click Ok to save the configuration. The following success message appears:

   

6. To delete a skill in the broadcast model, click on the delete icon corresponding to the skill to be deleted.

   

7. The following confirmation message appears. Click Ok to confirm the action.

   

Multiskill Configuration – Conductor-Follower Model

• Perform the following steps for MULTISKILL configuration using the Conductor-Follower Model:

1. In the Multi Skill Configuration page, select Conductor-Follower Model and click Continue.

   

2. In Conductor-Follower Multi Skill Configuration, select one Conductor Skill by using the Conductor Skill drop-down. Then add one or multiple Follower Skills by using Choose a Skill drop-down. Enter an Intent Name for the skill.

   

3. In the Conductor-Follower model skill, the intent names displayed in the following figures are added:

   

4. Click Save to configure the multi skill.

   

5. The following confirmation message appears:

   

6. To delete a skill in the Conductor Follower model, click on the delete icon corresponding to the skill to be deleted.

   

   Note:

   • In Conductor Follower model, we must add the Intent value while adding the Context Variable.
   • The Skill Intent Name must be unique in the Conductor Follower Model.
   • A skill that is configured in Multiskill cannot be deleted from the tenant workspace. The user must delete the skill from Multiskill configuration to delete it from the tenant workspace.
   • There must be at least one skill in the Multiskill configuration.
   • The skill name must be unique in all the models.
   • There must be Intent Value added while adding the Context variable in all the Multiskill models.
7. Once the Multiskill configuration is done, select the INTENT tab. If none of the skills is selected, then a pop up will be displayed saying select a skill.

   

1. Progress bar for loading the intents associated with the skill is visible as shown in the above figure.
2. If intents are present, the list of intents will be populated in the intent table. Refer the following figure:

   

3. To add new intent, perform the following steps:
4. Click on ADD Intent button under the Intent tab. The Add Intent screen appears:
5. The intent name can contain letters (in Unicode), numbers, underscores, hyphens, and periods. The name cannot consist of ‘..’ or any other string of only periods. Intent names cannot contain spaces and must not exceed 128 characters.

   

6. A success message appears, and the intent value appears in the intent table.

   

7. To delete an intent, click on the delete icon corresponding to the intent that you want to delete.

   

8. For handling utterances and intent recognition, variations provide a robust mechanism to train and recognize user conversations in natural language. Variations should be added so the intent recognition system can generalize.
9. New Variations can also be added to an intent by performing the below steps:
   1. Click on any value of an intent from the intent table under INTENT tab. A slide navigation slides in containing the Intent name at the top.

      

   2. Add a new variation and submit it by clicking on Add button.

      

   3. The newly added variation appears in the variation table as shown below.

   4. 

   5. You can delete the variation by clicking the delete icon corresponding to the variation.

      Intents can also be edited from the main intent page, click on edit icon for which the entity name user wants to edit it.

      

   6. A modal will open to edit the intent name, where you can make changes as needed.

      

   7. Now, select the ENTITY tab. If none of the skills are selected, then a pop up will be displayed saying select a skill.

      

   8. If the entities are present, the list of entities will be populated in the entity table.

   

10. The following columns are available:
    • Entity Name: Displays the list of Entities.
    • Fuzzy Match: Contains the toggle buttons that display the status of fuzzy logic of the entity.
    • Modified: Displays the time when the property of the entity was last changed.
    • Action: Contains the delete icon to delete the corresponding entity.
11. Perform the following steps to Add Entity.
    1. Click on the ADD Entity button present under ENTITY tab. The Add Entity screen appears:

       

       The entity name can contain letters (in Unicode), numbers, underscores, and hyphens. Do not include spaces in the name. The name cannot be longer than 64 characters. The name cannot be prefixed with ‘sys-’ as it is reserved for system entities.

    2. Add the Entity by clicking on ADD button. A success message appears:

       

       The newly added Entity appears in the Entity table.

       

    3. Values can be added to an Entity by performing the following steps:
       1. Click on the value of an entity under the ENTITY tab. A side navigation slides in containing the entity name at the top.

          

       2. Add a new entity value and submit it by clicking on Add button present there.
       3. Select the type from the dropdown. Two types are available:
          • Pattern option allows to put regular expressions for values.
          • Synonym option allows to match the text itself.

          

       4. The newly added entity value appears in the entity value table.
       5. To delete an entity value, click on delete icon corresponding to the entity value that you want to delete.
12. Take the following steps to add synonyms to the Entity:
    1. Click on any value of an entity in entity value table to add synonyms to it. A side navigation slides in containing the entity name at the top.

       

    2. Add a new synonym and submit it by clicking on Add button.

       

    3. The newly added synonym value appears in the entity value table.

       

    4. To delete a synonym, click on delete icon corresponding the synonym that you want to delete.
    5. Entity names can be edited by clicking on edit icon for which the entity name user wants to edit.

       

    6. A dialog box will open to edit the entity name:

       

    7. Perform the following steps to enable system entities:

       Pre-built entities allow users to extract and work on commonly used language references, such as dates, numbers, etc.

1. Click on the System Entities button under the ENTITY tab.

   

2. A window appears showing the available system entities. The console provides several system entities, which are common entities that you can use for any application. Enabling a system entity makes it possible to quickly populate your skill with training data that is common to many use cases.

   

3. Click on the toggle of any system entity to enable it. For example, the @sys-number system entity matches any numerical value, including whole numbers, decimal fractions, or even numbers written out as words.

   

4. You can now use the enabled entities in your dialog nodes directly.

   Contextual Entities can also be used with UCD console; these entities allow BigFix AEX to detect user context from the utterance. Contextual entities provide a faster way of recognizing newer entities without explicitly creating them every time.

   1. The value which is annotated in a variation of the intent will reflect in an entity by yellow color.

      

   2. Select DIALOG NODE tab. If none of the skills are selected, then a pop up will be displayed saying select a skill.

      

   3. The dialog node’s id and description (optional) appear on the card.

      

   Note:

   • While adding the Context variable, skill_change should be in true condition.

   • While configuring the Anything else node, the anything_else condition must be ‘True’.

   

5. To add new nodes, click on ADD NODE button under the DIALOG NODES tab.

   

6. To see the child nodes associated with the Node, click on Expand More.

   

7. The associated child nodes open.

   

8. On clicking of Expand Less, the child nodes collapse.
9. The following actions can be performed on each available node by clicking on the kebab menu (three dots menu) placed on each node:
   • Add Child Node: A new node will be added as child.
   • Add Node Above: A new node will be added above the node.
   • Add Node Below: A new node will be added below the node.
   • Delete: Deletes the Node
   • Move: Move within Nodes

   Moving a dialog node that you have created can be moved anywhere in the dialog tree. You might want to move a previously created node to another area of the flow to change the conversation. You can move nodes to become siblings or peers in another branch.

   On the node if you want to move, click 3 dots icon, and then select Move.

   

10. Select a target node that is in the tree near where you want to move this node. where user will find three options i.e., As child node, below node and above node as shown in fig.

    

    • Jump to: Jumps to the following options as defined:
      • Response
      • Assistant Recognize
      • Wait for User Input

    

11. Jump to: To perform “Jump to” on a node, click on type of Jump to selector i.e. Wait for User Input, Assistance Recognize, or Response. Select destination node by clicking on moving point (MP icon )

    

    

    The source node of Jump to option will be labeled with Jump to (destination node id) after a pop-up message stating that Jump to has been applied successfully on Watson.

    

12. To delete the ‘Jump to’, click on delete icon present next to Jump to (destination node id) label.
13. Once Jump to has been initiated but needed to be cancelled before selecting a destination node; in that case click on cross icon present on the right bottom of the console.

    

14. By clicking on Open placed on each node, you can view the form where you can perform multiple actions like view or modify the settings of the node, add context, and add Response etc.

    

    

15. Click on the setting icon to check the settings of form. The following screen appears:

    

16. To set digressions, click on that toggle button of digression and to set multi condition response, click on toggle button of MCR.

    If MCR is enabled, the form shows the fields specific to MCR form.

17. Note:

    When dialog node form is opened for first time, it shows standard form with fields and buttons i.e., node name, condition, add context and response type.

18. Enter Node Name and Condition.
19. Click on ADD CONTEXT and ADD RESPONSE to add new context and response respectively. Refer the image below to understand how to add conditions, context variables and Responses.

    

20. The Intent name must be prefixed with # and the Entity name must be prefixed with @ to match the condition. You can also use logical operators such as &&, || and! in conditions. To save any entity/intent matched as a context variable use the format given in the picture above.
21. The name of context variable can contain any upper and lowercase alphabetic characters, numeric characters (0-9), and underscores. You can include other characters, such as periods and hyphens, in the name. However, if you do so, then you must specify the shorthand syntax $(variable-name) every time you subsequently reference the variable.
22. The value of context variable can be any text that you want, but if you want to capture an intent/entity from user input, use the following expression syntax to capture it:
    <?@entity_name?>.
23. Clicking anywhere in the form submits the fields.
24. Addition or deletion will be enabled once the fields are submitted successfully. You can add more responses by clicking on the ADD RESPONSE button:

    

25. If there is more than one response in a dialog you can choose one of the three settings that the console provides.

    

    • Sequential: The system returns the first response variation the first time the dialog node is triggered, the second response variation the second time the node is triggered, and so on, in the same order as you define the variations in the node.
    • Random: The system randomly selects a text string from the variations list the first time the dialog node is triggered, and randomly selects another variation the next time, but without repeating the same text string consecutively.
    • Multiline: When the response is shown to the user, both response variations are displayed, one on each line. For example, if you have 3 response variations: Hi, Hi there, and Hi there friend. The multiline response would be as follows:

      Hi

      Hi there

      Hi there friend.

26. When MCR is enabled, the MCR form is displayed with fields and button i.e., node Title, Condition and an “Add Multicondition” button. The pop-up message will be displayed stating that MCR is enabled. By default, the Add Multicondition button is disabled.

    

Perform the following steps to configure multicondition:

1. The values are fetched and auto populated in the respective fields and ADD MULTICONDITION button will be enabled.
2. Click on Add Multicondition button to add new multiconditions.

   

3. All these fields on MCR form get submitted on focus out. Addition or deletion options are enabled once the fields are submitted successfully.
4. Click on edit icon associated with each multicondition to add other values to that multicondition. A screen appears with fields i.e., Condition, Add Context, Response Type with a Submit button.

   

   The condition field inside the edit screen will have same value as the multicondition which can be edited.

5. Click on ADD CONTEXT and ADD RESPONSE to add new context and response of the multicondition respectively.
6. All these fields on MCR modal form will get submitted by clicking on the Submit button below.
7. A success message appears once the form is submitted successfully. Addition, Deletion or Edit of new multicondition will be enabled once the edited details are submitted successfully.

   

8. Close the form by clicking outside of the form.

   You can chat and test the use cases configured to see if they are responding as per the requirement or not. Perform the following steps to do so:

   1. Click on TRY IT on the UCD Console under Dialog Nodes tab. A chat console appears.

      

   2. The Try it out panel slides in as shown in the following image. Use the text area at the bottom to enter your query and click on TEST to send it.

      

   3. The response looks like follows:

      

   4. Here you can see the console detected the entity @sys-person:Lisa. Click on the blue location pin to open the dialog node that was triggered by your input.

      

   5. Click on the CLEAR button on the top to remove any previous chat from the Try it out panel.
   6. Click on the MANAGE CONTEXT button to create or make changes to context variables.

      

   7. Enter the variable name for context and then click on the plus (+) icon to add value as desired.

      

   8. If any intents are matched, it will show up as follows. Here the intent #turn_off is detected:

      

   9. Clicking on the eye icon opens the list where it shows all the intents and the confidence score of the intent for the input query.

      

      The Analyse feature lets the user upload the usecases and its variations and verify what is the predicted intent and usecase for them.

      • The ANALYSE button is present under Intent, Entity and Dialog Nodes tab as shown in below figure:

        

      • When ANALYSE button is clicked the following screen pops up:

        

      • Clicking on DOWNLOAD SAMPLE CSV downloads a csv file which is a reference template to upload the usecase variations file.

        

      • Once the file is uploaded the READ button gets enabled.

        

      • When we click on Read button, the data is uploaded from the csv file to the table on the screen and the RUN TEST button gets enabled. Refer the below screenshots:

        

      • If the CSV is in improper format the following error alert appears as shown in the figure below:

        

      • When there is no data in the csv file and it is uploaded, the Read button gets disabled, and the table shows the text “No data uploaded”.

        

      • Once the rows from CSV are read and the data on the screen is uploaded and displayed, the RUN TEST button is enabled. On the click of RUN TEST button, the analysis of variations for all the data rows starts.
      • The screen displays the progress of the analysis for all the rows of the table and the RUN TEST button is disabled. Once it’s completed the progress bar disappears.
      • On the upload of a different file the READ button gets enabled.

        

        

        Note:

        Refer to Appendix A to get the list of Rate Limits of the APIs used in the Use Case Design Console.