Speech analysis

The robots are able to identify many words (it is not a rule, but frequently happens), and so use this ability. Consider an example. Our computing power solve the tasks regarding a moment of inactivity by implementation of banking programs. There are 40-50 thousand of calls where a machine finds the pauses, analyses everything happened before and after and does conclusions about a process based on this data. Here is a case you can get as a result: for your information, on the Monday morning an operator lingered for an answer regarding the opportunities to take a loan in a bank on bail. Or there are delays when, for instance, sportsman – the client of insurance company – is going to unpopular country and wants to arrange interesting insurance, but an operator can make him wait few minutes before he/she finds the needed information in his/her interface. By the same principle, it is possible to detect the breaks of the interfaces themselves.

Obviously, that possibility to recognize the words helps to track the way the operators communicate with the clients: greeting manner, using fillers and obscene expressions, tactfulness in conversation etc. Also there is ability to know the clients’ opinion regarding a call: rejectionы like “I don’t need, thanks” followed by voiced proposals, positive and negative reviews and so on. Detailed on MoveUp

You can meet emotional analysis too. It is very useful system even though it is still on its way to a perfect precision. Many banks experience such decisions especially in the collection services. For instance, if during a call is a long pause or the interlocutors begin an argue, or start to swear – supervisor is notified immediately.