Ask: Open

To ask the user an open question with no predefined choices

The Ask: Open node variant is used to ask an open question for which no predefined options are needed. Like for instance asking for the name of a company, someone's opinion, or their reason for calling.

Ask: open

Node properties

Property
Description
Explanation

Question

The question to ask

Assign to

To assign the result to a variable in the conversation

Remember

To also store the variable in the user object

The variable name in the Assign to will be used to store the value in the user object.

Record audio ☎

On phone channels there is an option to record the audio

Assign audio URL to

If record audio is toggled, the URL to the audio file will be stored in this variable

Only displayed and mandatory when Record audio is toggled.

Capture patiently on phone

To allow for longer voice input. This is done by waiting two times longer before stopping the recording. This can be overruled by the user by pressing the hash (#) key.

The recording will be fed to the LLM provider to transcribe, which results in slightly better transcription than without this patiently option.

Check answer relevance

To check whether or not the answer is relevant to the question.

The question and answer will be fed to the LLM provider to see if the answer has relevance to the question. If not, the answer is rejected, and the user will be asked again.

If remember is toggled, this node will also set a user tag using the assign to value. Don't forget to manually create the page Tags

When used on phone channel

On phone channels the option to capture patiently is useful when asking for longer text input, like when asking about the reason of a call, or when asking the user to speak in a complete product order with quantities and product names.

The capture patiently option will listen two times longer for user input. This way, users can take their time to speak without being cut off halfway. They can always use the hash (#) key on their device to stop recording and continue. Make sure to explain this in your question text.

If the capture patiently option is selected, the transcription will be more precise because it will be transcribed by an LLM provider before storing it in the assign to variable. This might cause a slight deviation between the transcription presented in the conversation transcript versus what is stored in the assign to variable.

If the LLM provider takes too long to transcribe, the user will get a wait text in the meantime: "One moment please" to let the user know their input is still being processed.

If using capture patiently, the transcribed text that will be stored in the assign to variable might differ from what is shown in the conversation view since the transcription is done using a different provider.

Resources

Last updated

Was this helpful?