i'd probably test a small set of conversational queries manually. its not perfect but it gives you better idea of what assistants are actually returning.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.