WorksheetsODA Review Session
Total questions: 32
Worksheet time: 8hrs 0mins
You want your skill to transfer conversations over to Oracle Service Cloud customer service representatives. Which type of channel do you create to enable the skill to do this?
Applications
Agent Integrations
Users
System
want to save some user input, such as the type of pizza a particular user last ordered, so that it's available the next time that user starts a conversation. Which type of variable should you use to persist values across multiple invocations of the conversation?
skill variables
profile variables
context variables
user variables
Which variable type is automatically set with values from the chat client, such as locale and timezone offset?
System Variables
System.config variables
User variables
Profile variables
Which two statements are true for the system.webview component?
Parameters passed from the skill to the web application are accessible through JavaScript.
Parameters passed from the skill to the web application cannot be accessed through JavaScript.
Property names in the web application response payload must match with the variable names in the dialog flow.
The webview.onDone parameter is automatically added to the payload and passes the skill's callback URL property to the web application.
When registering web applications in a skill, bot designers can see the data structure returned by the app
What statement correctly describes the Authentication Service on Oracle Digital Assistant?
The Authentication Service holds the identity provider configuration that is used at run time in Oracle Digital Assistant to retrieve an access token that authorizes
REST service calls.
The Authentication Service authenticates users to Oracle Identity Cloud Service
It provides customizable login screens that are displayed in the context of a user-bot conversation.
The Authentication Service allows bot designers to configure a custom webhook to authenticate and authorize users using the System.OauthAccountLink
component
What is the output of this code?
All the three values - "One." "Two." and "Three." will be printed.
The code will fail to validate because |- is not a valid symbol.
Only first value - "One" will be printed.
The code will run into an infinite loop.
You install Oracle Bots Node SDK from GitHub to develop a new custom component service. Which command, when issued on a command line or terminal window, creates a new custom component service project in the current directory?
bots-node-sdk init
bots-node-sdk service init
bots-node-sdk service
bots-node-sdk npm install
You want the flow to navigate to the cancel transition immediately after the maximum number of failed attempts are exceeded in the System.ResolveEntities Components. Which option must you use?
Set cancelPolicy to "true" .
There is no such option in system.ResolveEntities
Set cancelPolicy to "immediate".
Set cancelPolicy to "lastEntity" .
Within your digital assistant, you notice that the user input "tell me my balance" immediately initiates the Banking skill. However, it does not offer the user the option to consider that the request could be handled by the Retail skill, which also offers the ability to check the balance in your retail account. How should you ensure that both the banking and retail skills are considered in this case?
Lower the Candidate Skills Confidence Threshold in the digital assistant.
Lower the Confidence Threshold in the Retail skill.
Raise the Candidate Skills Confidence Threshold in the digital assistant.
Lower the Confidence Threshold in the Banking skill.
How do you declare a context variable for an entity?
Set the variable type to "map" and reference the value by the entity name.
Set the variable type to "entity".
Set the variable type to "nlpresult".
Set the variable type to the same name as the entity.
You are advised to implement an 80/20 split with training and test utterances. This means that 80% of new utterances harvested from the conversation logs should be used as intent training utterances and 20% for testing. Why is this a good practice?
Batch testing works more efficiently when there is a ratio of one test utterance for every five training utterances.
By keeping 20% for testing, you are able to test the model with data on which it has not been specifically trained.
Adding 100% of user phrases to the intent would overload the model.
By performing an 80/20 split, you are randomizing which data is added to the utterances.
Which two features of Oracle Digital Assistant should you use to allow a skill to specifically classify the user message "What's my balance in savings?" and to identify the type of account?
Dialog flow with system.text component to read the user input
an intent that is been trained with utterances such as "check balance", "what is my current balance?", or "how much money do I have?".
an input form rendered by a webview that is hosted within a digital assistant.
An entity that defines account types (with values such as "checkings" or "savings")
Which two component can be used in combination with composite bag entities to auto-generate skill responses and flows from definitions saved in bag items?
System.Text
System.CommonResponse
System.ResolveEntities
System.MatchEntity
For Agent Integration, you want the bot-user conversation history to become available to the called human agent. Select the configuration option you make this happen.
In the skill settings, switch the Enable Insights option to "On".
Set the "convHistory" variable in the System.Agentinitiation component.
Set a custom property on the Service Cloud instance accessed by Oracle Digital Assistant.
This is controlled from Service Cloud and has to be turned on by setting a custom property.
You have gone through a number of testing iterations of your customer's skill that comprises 10 intents. But you find that generally the best you can get is a confidence score of 96%, even when the user phrase is identical to one of your training utterances. What should you recommend to your customer regarding this intent confidence score?
It is not always possible to achieve 100% confidence and adding more utterances may not help the problem. Therefore, do not make further changes to the skill if it is performing to your expectations.
Keep iterating on user testing and add more training utterances until you can achieve a confidence level of 100% on your user input.
The highest possible confidence with 10 intents is 10% (100% divided by the number of intents). So, no further changes to the skill are required.
For every verb in your training utterances, ensure you add a version of the utterance which also covers the past, present, and future tense of the verb.
Add more utterances to the unresolvedlntent.
Which statement is FALSE regarding out-of-order messages?
Chatbots don't control the user's input and, therefore, cannot prevent users from selecting out-ofscope
actions.
Out-of-order messages occur when a user scroll the conversation history displayed in the messenger
client and selects that is no longer in scope for the current conversation.
Out-of-order messages are not handled by default. Designers must define out-of-order message
handlers at the skill level.
Dialog flow navigation continues with the state referenced in the out-of-order-message action.
To prepare the remote application launch, the System, Webview component calls an intermediary service that prepare the remote web application call. The codesnippet is given below:
At run time, the system.Webview component sends the intermediary service a POST request. Which three options are true for the POST payload generated based on the code?
The key names match the name of the dialog flow variables configured in the "sourceVariableList" property of the system.webview component.
Bot designer is responsible to manually add the webview.onDone parameter to the payload.
The webview.onDone parameter is automatically added to the payload and it passes the skill's callback URL property to the web application
There is no such webview.onDone property passed to the payload
The POST payload is a JSON object that contains an array of key-value pairs.
In your conversation flow, you want to make sure that users always see a message, even when there is no data to display. To implement this, you decide to use a system.SetVariable component that verifies that the variable mydata contains a value and, if it does sets the value of the displayVar variable to the value of mydata. If no value is specified for mydata, then displayVar is set to the string ‘No Data.’
Which two BotML with Apache FreeMarker examples implement this requirement?
You have a skill for ordering pizzas. In your experience, you have found that 95% of your customers want a regular crust and just 5% prefer gluten-free crust. Because so few people are likely to want a gluten-free crust, you don't want to prompt for that option, but you do want to make it available. Assuming you already have a composite bag entity that contains a PizzaCrust entity that has the values "regular" and "gluten free", what would be the simplest way to have your skill enable a user to order a pizza with a gluten-free crust without the skill prompting the user for that option?
In the composite bag, set the PizzaCrust entity item's Prompt for Value property to False. Then, once the the composite bag is resolved, check if the PizzaCrust entity has a value. If it doesn't, set its value to "regular". Don't set a prompt for the PizzaCrust entity item and set the Out of Order Extraction property to True.
For the PizzaCrust entity item, set the Extract With property to reference an entity with the single value "gluten free".
Don't set a prompt for the PizzaCrust entity item.
Which two statements about using the OAuth2 client credential grant type in Oracle Digital Assistant are correct?
The user must be logged in to an identity service provider using his or her username and password to obtain a client credential access token
The OAuth2 client credential grant type does not require a bot user to authenticate and instead uses a shared clientld and secret.
You use the Oracle Digital Assistant Authentication Service with the system. outh2client component to obtain a client credential access token.
The OAuth2 client credential grant type uses the system. Webview built-in component to perform social media login.
The client credential grant type requires the system.OAuth2AccountLink component to obtain a token that propagates a user's identity between distributed systems.
Skill to output the size of the pizza that was ordered. Which FreeMarker operation must you use
“You ordered a ${size[0]} pizza." 3)
“You ordered a ${size.value} pizza.
“You ordered a ${size} pizza." :)
"You ordered a ${size.string} pizza."
Which two statements about message translation in a skill are true?
--If auto-translation is enabled and a component has its translate property set to false, then the component output message or level will not get auto-translated to the detected user languages.
--A system.Output component that reads its text message from a resource bundle does not require auto-translation or its translate property set to true to display translated.
DetectLanguage state in a dialog flow causes an exception for components that read their output message from bundle.
For the System.Translateinput component to work, it requires a previously executed system.DetectLanguage component state.
A missing system
Enabling auto-translation in a dialog flow does not translate the user input message.
You are writing a dialog flow and you are at the end of the conversation. For the final state, you are not sure whether to use a return transition use a next transition that goes to the start of the dialog flow. Which statement is true?
The next and return transitions are the same and you can use them interchangeably.
The next transition can only be used for forward navigation in the flow.
The next transition automatically clears variables on navigation, whereas return doesn't.
The return transition goes to the start of the dialog flow and resets all variables, whereas next simply navigates
Which statement is true regarding the effect of context pinning on routing?
The next utterance, the router pins the conversation to the skill
The thresholds that determine whether context pinning will occur or not set at the skill level.
If the user is pinned to a skill and then enters something that doesn’t relate to that skill, the router automatically checks other skills for a m
If the input includes an implicit invocation for a skill, the router pins the conversation to that skill.
For user input that includes an explicit invocation for a skill, but not intent-related utterance, the router pins the conversation to the skill
Which three options are true for the system, entityToResolve variable?
${system.entityToResolve.value.userinput} returns the text entered by the user.
${system.entityToResolve.value.resolvingField) returns the text entered by the user.
The system.entityToResolve variable can be referenced from withinthe system ResolveEntities and system commonResponse components to display, for example, information about the entity that has been resolved.
The system.entityToResolve variable tracks an entity value, that is, as you resolve entities in the composite bag, it references the current entity resolved.
System.entityToResolve can reference the resolve value of all entity values defined within the skill.
How does a digital assistant determine which skill to route user input to?
It routes to the skill that is activated most often.
It evaluates all of the user input against the individual training models for each skill, and then it routes to the intent with the highest confidence score.
It evaluates all of the user input against one consolidated training model that combines all of the intents, and then it routes to the intent with the highest
confidence score.
it evaluates against the training models for the candidate skills and resolves to the intent with the highest confidence score.
It evaluates all of the user input against one consolidated training model based on the utterances for each skill to determine the candidate skill
Which THREE of the following are capabilities of digital assistants?
Greet users
Translate responses to the users language
Enable users to interact with multiple skills in one session
Invoke other digital assistant
Listen for requests to exit the digital assistant
You have a Pizzeria Skill, where you have set the Confidence Threshold to .6 and the Confidence Win Margin to .2. A user enters input that results in confidence scores of 79% for Order Salad, 69% for Order Pizza, 61% for Order Pasta, and 17% for Cancel Order. Based on those settings and confidence scores, which intent(s) are matched to the input?
Order Salad
Order Salad and Order Pizza
Order Salad, Order Pizza, and Order Pasta
Cancel Order
None
In OBotML, you can branch the flow of conversation depending on the users input by using which of these mechanisms?
If/then/else statements defined globally
For/case statements defined globally
The actions node under the transitions node within a state
The next node under the transitions node within a state
What is the search order for out-of-order message handlers?
1 - system level handler,
2 - current dialog flow state,
3 - global default transition
1 - current dialog flow state,
2 - system level handler,
3 - global default transition
1 - current dialog flow state,
2 - global default transition,
3 - system level handler
As defined in OBotML
You have a custom component service that you want to share among multiple skills without exposing multiple copies of the custom component service. Which TWO of the following ways can you achieve this?
Host the service in an embedded container in the digital assistant.
Host the service on Oracle Mobile Hub
Host the service on Oracle Content and Experience Cloud
Host the service on a private node.js server
In a digital assistant, the response to a user's input is sometimes a choice between multiple intents. Which of the following can you do to increase the likelihood that a user will only be presented with the top-scoring intent?
Increase the Explicit Invocation Confidence Threshold
Increase the Interrupt Prompt Confidence Threshold
Increase the Built-in System Intent Confidence Threshold
Lower the Confidence Win Margin
