So ... thinking about this ... your supervision offerings historically have focused on Exchange Online for FINRA. The new updates above point to a broadened focus, such as the inappropriate language and legal exposure use cases. These could be implemented in a couple of ways in Office 365 - DLP could do it, or Supervision could do it, or probably even Office 365 Cloud App Security (the logic model is there).
For example, Unified DLP could be set to look for inappropriate language, documents with particular retention labels, or documents with specific words or phrases (doing a lexicon match, etc.). Why did you strategically decide to do this in Supervision rather than DLP (granted that DLP does signal "data loss" so would be fine for the confidential projects use case, but not so much for language concerns)?
We all sit in the outer ring and are affected by the choices made within the mothership. I guess I'm interested in the background thinking.