Lex-bot utterence have only "Yes, yeah, yup" but the bot work/starts/invoke with other words as well.

0

Hi Team,

  1. I have created a demo lexbot which have two intents, the default fallback intent and one other "Test" intent. the Test intent have around 6 slots. in this intent I have added some utterances like "Yes, yeah, yup" only, and I have added the question (If you are an employee please say yes) in the Amazon Connect get customer input block where I am invoking the bot. but the Bot works with other words as well, like no, car, bus etc, => How Can I control the bot so that bot go to fallbackintent with words other than "Yes, yeah, yup"??

  2. Also I have a built in slot type employeeNumber which have a regex [0-9]{4,6}, But while interaction during voice, it accepts less than four digits as well like 12, 32, 334, etc. with chat it works fine, only have problem with voice. => How can I solved this issue?

thanks in advance team.

1 Answer
1
Accepted Answer

Hello,

I did not get question 1, do you want to create more utterances that will send you to failback ?

responding to question 2, you could define a custom slot validation using AWS Lambda. This function should check if the input has 4 to 6 digits and reject inputs that don't meet this criterion.

profile picture
EXPERT
answered 18 days ago
  • thanks Giovanni, for the second second answer, let me rephrase the first question. currently the bot have only two intents, the fallback and another test intent, in the test intent I have added utterances like, "Yes, yeah, yup" to get the bot started, when I ask the question in the prompts like, If you are an employee please say yes, here if the caller says yes, then the bot will go the the slot and will ask "Whats your name", since the caller is an employee. but when the caller says some other words like no, car, bus etc, it still ask for "Whats your name" rather than go to the fallback intent. thanks

  • Add more utterances to the Test intent to better train the classifier. For example, add responses like "no", "not an employee", etc. This will help the bot distinguish between positive and negative answers.

    Alternatively you can use intent contexts to control the flow. For example, set the employee check intent as the current context after a "yes" response. Only allow the name slot intent when this context is active.

  • It worked like that. I created another intent and added all the possible responses to that in the utterances. No it works fine. regarding the first question, I have to use lambda for custom validation.

    thanks Giovanni, for the quick responses.

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions