FAQ API
Hi I am trying to use the FAQ API to get more control over certain responses. However, I cannot find a way to get more control over the sentiment. For example with chunks I can control relevance with the scoring whereas with FAQs I don't seem to have that option. If I create for example the following FAQ I get these responses to the questions asked:
FAQ Quetion: Email?
FAQ Answer: xxxx@xxx.com
when the users asks: Email?
response: xxxx@xxx.com
but when the user asks, Can I contact you via email?
response: xxxx@xxx.com
in the second case which is not a "close" match I am still getting the FAQ response resulting in inaccuracies. Is there a way to have a bit more control over the confidence score of the FAQs?
FAQ Quetion: Email?
FAQ Answer: xxxx@xxx.com
when the users asks: Email?
response: xxxx@xxx.com
but when the user asks, Can I contact you via email?
response: xxxx@xxx.com
in the second case which is not a "close" match I am still getting the FAQ response resulting in inaccuracies. Is there a way to have a bit more control over the confidence score of the FAQs?
