Oracle Digital Assistant supports several messenger-specific channel types, such as iOS, Android, Web, and Facebook. If your messenger client is not supported, what is the simplest way to connect your skill or digital assistant to the client?

• A.You can't connect to a unsupported messenger client. You canonly connect channels to the supported clients.
• B.Implement the channel using a Webhook.
• C.Ensure that your skill only outputs text and not Images or visual components such as cards. Then use the standard Web channel, which you can use for text-only conversations.
• D.Use a custom component to extend a supported channel type.

Comment: *

Name: *

Email: *

Verification: *

Which is NOT used to tune routing behavior?

• A.the built-in systemintent confidence threshold
• B.the classifier's F1-score
• C.the confidence win margin
• D.candidate skills' confidence thresholds

Comment: *

Name: *

Email: *

Verification: *

Which three statements are true about composite bag entities?

• A.The composite bag will always enforce that every entity has a valid value before allowing the conversation to move on to the next state in the dialog flow.
• B.When you add entities to the composite bag, you can control how they get extracted in related to other entities and when they are prompted for.
• C.They define a business domain object as a collection of related system entities and custom entities.
• D.You need to create a separate composite bag to handle nonentity types such as strings. Locations, andattachments.
• E.The composite bag can resolve all entity values using only a single state in the dialog flow.

Comment: *

Name: *

Email: *

Verification: *

As per Oracle's recommendation,which is the best practice regarding conversational design?

• A.Ask users open-ended questions such as "how can I help you?"
• B.Use quick reply buttons (as opposed to natural language inputs) as much as possible.
• C.Ensure that capabilities of the bot f the things that it can and can't do") are clear and discoverable.
• D.To account for possible mistakes, make it clear to users that the bot is still learning.

Comment: *

Name: *

Email: *

Verification: *

You are advised to implement an 80/20split 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?

• A.By performing an 80/20 split, you are randomizing which data is added to the utterances.
• B.Batch testing works more efficiently when there is a ratio of one test utterance for every five training utterances.
• C.Adding 100% of user phrases to the intentwould overload the model.
• D.By keeping 20%for testing, you are able to test the model with data on which it has not been specifically trained.

Comment: *

Name: *

Email: *

Verification: *