IBM software is able to understand the emotions of users – Info Online

The research lab IBM in Israel began testing a customer service software able to understand feelings. Still not be named, the system can recognize basic emotions – like anger -. A user that comes into contact with the company through an automated chat or even by Twitter

According to a report in the Wall Street Journal, the platform “learns” to interpret emotions from data obtained from e-mails and postings on the microblogging network. In lot of information, it is able to identify words, grammatical structures of the sentences and contexts. Until some more specific features such as the use of exclamations, caps, emojis and abbreviations (LOL, WTF and OMG, for example) and accelerated typing are understood by the software – and everything is then classified according to the emotions they represent.

It is from this that the software can understand what you are feeling and “adapt their approach and even their language,” according to WSJ. If a client is losing it or beginning to take explanations, the system can ask apologize and detail the actions it will take to resolve the issue. And if the consumer is already without some patience – and writing in Caps Lock or sending profanity – the platform may end up redirecting it faster to a flesh attendant

The goal, therefore, no. It involves replacing human attendants, who are still considered essential by one of IBM scientists heard by the report. But only for now. “Our goal is to show that we have reached at least the same level of satisfaction with the consumer you would with a good clerk,” said the newspaper Dr. David Konopnicki, responsible for the project. For him, the full exchange of people by machines is only a matter of time and adaptation of the same customers.

The first prototypes of the software should be released next month for some IBM client companies. They should add this “sentimental layer” to its today insensitive automated chat systems, which appear in the corner when the consumer accesses the company’s website. The software should only gain support voice calls later, but you may soon have the opportunity to talk into text with a robot that is almost a person.