A project I'm going to be working on will be utilizing Amazon Lex exclusively for the purposes of intent matching. Looking at the docs for PostText, I can see that we're required to send a userId to Lex in order to use this method. For our use-case we don't really care about the conversational capabilities of Lex, so we have a few options as to how we could approach the userId field:
Just use a global userId that every single lex query would have
Generate a random userId each time we query lex
Generate a userId per specific user of our system and use that each time that user queries lex
My gut feeling is that approach 3 is probably the "correct" way of doing this but maintaining an id for our users may be cumbersome if we don't have to do it. Between these three approaches what makes the most sense for our use-case? Thanks!