I wonder what approach is used to feed the Pulse system. Not so much the data flow done in real time (we already know that a strength of the Verb tools is the capacity to follow in real time what happens to create reports, allow immediate responses, feed other CRM software etc) but rather the data and method used to train of the recommendation system.

As you say, I assume it is AI (though it might be complemented with some expert-based rules, in particular for brand new customers and/or markets with limited data) but I do wonder what approach they follow

My guess would be some dynamic supervised machine learning that will « uncover » the actual best practices : what do the best sellers (those that do have the best figures) do in practice in various situations (timing, type of customer, type of messages…). Then based on that suggest similar actions to the other sellers at the right time.

After all, Rory often mentioned that not everybody is a great sales person and that one of the goals of Verb tools is to allow anyone to reach great results. It applies to the video themselves but also to the best way to communicate with a client