Interacting with a virtual assistant can quickly become boring if the voice is too robotic. This is why Amazon has just launched “Alexa Emotions and Speaking Styles”, two new features that give more relief to the voice coming out of the connected speaker Echo.
Thus, Alexa now has two emotional modes, namely “happy/excited” or “sad/empathic”. The first could, for example, can enter into use if “the customer answers a question correctly or wins part of a game”. The second would be appropriate if “the customer requests sports information and his favorite team has lost”.
The “Speaking Styles”, on their side, are not meant to convey an emotion, but to bring more naturalness. Two modes are available today: information and music. With them, the voice of Alexa can take the intonations of a newscaster presenter or a radio operator.
For Amazon, the goal is obviously to increase interactions with its connected speakers. According to the manufacturer, these new ways of speaking increase user satisfaction by at least 30%. Technologically, they use machine learning algorithms “Neural Text-to-Speech” (NTTS), which are able to synthesize voices depending on a universal vocoder.
For now, these new features are only available in English. But it is very likely that they will reach other languages too.