Corporates in censor-land social media are often semiautomated, and the human component is often plural and hierarchical, or even outsourced. I don't know if bot/not bot really captures the variability in note-sources.

What decisions will people make based on this information?