Enhance uses state-of-the-art machine learning to continually refine the recommendation engine that powers our suggested hashtags
When you use Enhance, RiteForge analyzes your post/Tweet/update in its current state and decides what it is about and then, either makes 1-2 words within the post hashtags (one green, one blue, as possible) or if this is not optimal, adds one green and one blue hashtag to the end. When you either accept or edit those hashtags, you actually improve our suggestion engine slightly. Many users doing this over time makes us better - and lets us deliver a more refined experience with better suggestions over time. Even your decisions on reducing the post text, adding images, emoticons (emoji), and other things you do to optimize teach the machine what is really optimal.
Here's how I took the Enhance green and blue hashtags, but added an image for "Ishi," the last of his tribe in California, by clicking in ishi, then the Free Images to Add tool, then changing from Flickr to Imgur as image source, and in two quick clicks, the perfect image for Ishi was added. I also changed the social network accounts to schedule to:
Here's more on machine learning, in general.
Auto-Emojify also uses machine learning.