NEW
Font size
WorksheetsDesigning for Conversational AI
Total questions: 10
Worksheet time: 5mins
Which of the following is NOT a purpose of entity tagging?
To train and test the virtual agent
To create relationships between words with similar meaning
To separate an utterance into predefined categories such as names, places, and organizations
To cleanse the data of unnecessary words
What is the difference between NLU and NLP?
NLU is for parsing texts while NLP has to do with speech.
NLU is a subset of NLP that focuses on reading comprehension.
NLU is for voice-based virtual agents while NLP is for text-based virtual agents.
NLU and NLP are synonymous.
Which of the following is NOT a best practice for conversation design?
Always use visuals
Use the active voice over the passive voice
Avoid using acronyms or jargon
Provide suggestions to guide the user
During a dialog, a user has the choice to select from a list of complex options.
Which of the following design elements would work best to describe each option?
Text-based answer describing each option
Tooltips
Link to FAQ page
Nothing - let users ask questions if they are confused
What is a context variable?
Temporary information that is stored during a conversation
User data that guides the tone the chatbot should use
Vocabulary related to the intent
A type of slot
When it comes to live agent handover, which of the following is NOT a best practice?
Give an accurate estimate of the wait time for transfer
Apologize if appropriate
Provide chat history to the live agent
Switch to a different channel
What is a fallback intent?
An intent that is triggered when the virtual agent does not recognize the utterance
An intent that is pre-defined in the NLP engine
An intent that initiates small talk
An intent that the user can invoke for live agent handover
What are control intents used for?
To help the virtual agent understand variations of a user’s question
To train the virtual agent to handle additional use cases
To allow a user to engage in small talk with the virtual agent
To help a user navigate through the conversation
Which of the following is NOT a reason to use SSML?
To show a voicebot how to pronounce nonstandard words
To make a voicebot emphasize specific words in a response
To train a voicebot to recognize users with different accents
To vary the rate and pitch of a voicebot’s speech
Which of the following is NOT one of the 5 response types?
Personalized Direct Response
Generic Dialog
Personalized Dialog
Targeted Direct Response
