It is critical in analyzing user utternaces to be able distinguish customer utterances vs button pushes in NLP Insights exports. See attached. The utterances in green are actually typed by the customer. The red are from a button pushes. Requirements

1. Distinguish customer utterances, buttons and possibly other text that would show up in this field. Do we need to change tags? Code?

2. We need to be able to do this at scale. Optimally we can temporarily add something to the session or message tags, or maybe set something in the traits until we get some reporting VIA API’s.