However, running rasa shell with debug, I saw that while the firstname is asked and captured properly into the slow, the zipcode is not. the question and answers are:
bot ask: great, we're almost there. what is your zipcode?
Hi @William8Work can you confirm that your entity extraction works outside the form? If you evaluated the message “my zip code is 98075” is your model able to accurately extract the zip code?
Very good question. Thank you @m.vielkind. The zipcode intent was missing from nlu so after I added it in, it seems able to extract that now. However, because I added another questions to ask for phone number, the NLU extracted both phone number and the zipcode from the same answer of phone number.
I use the following, @m.vielkind. However, the issue is that when the customer answers the phone question with a phone number, the bot fills both phone and zip code slot at the same time, and skipping the zipcode question.
nlu:
- intent: infom_zipcode
examples: |
- it is [12345](zipcode)
- my zipcode is [98101](zipcode)
- the zipcode is [91075](zipcode)
- zip code is [88045](zipcode)
- it's [98744](zipcode)
- intent: inform_phonenumber
examples: |
- my number is [2063303311](phonenumber)
- phone number is [411 3941112](phonenumber)
- cell is [415 303 2268](phonenumber)
- regex: zipcode
examples: |
- [0-9]{5}
- regex: phonenumber
examples: |
- ^(\+\d{1,2}\s)?\(?\d{3}\)?[\s.-]\d{3}[\s.-]\d{4}$