At the beginning there is a data set with emails. In an "initialization run" all emails are automatically assigned to a so-called categories. By doing so, a number of so-called "hot topics" emerge, which represent recurring business transactions.
However, Mailytica does not just recognize the categories. In addition, the data set analyzes where the emails were forwarded within the individual category in the past. This means that the rules are automatically detected and defined based on the dataset. Of course, these routing rules can be checked at any time and adjusted manually if necessary.
After the initialization run, Mailytica knows the "hot topics", the recurring business transactions, and can automatically forward new incoming emails to the desired locations according to the routing rules. Emails that can not be clearly classified can be forwarded manually at any time. With each manual forwarding Mailytica learns and apply this new knowledge to all future emails. Mailytica thus is continuously trained in new and old categories and can continuously increase the routing rates.
AI Email Routing is an iterative process - the application is constantly learning through Artificial Intelligence. The breadth and quality of the decisions thus grow day by day.