Facebook that runs through Messenger’s first chatbot with KLM, Microsoft, which up to the plate on this same segment during the last developer conference Build (with more or less success), Slack or Telegram embarking on the market bots programmable by everyone; the hour is to Artificial Intelligence on mobile and servers.
In 10 years, 50% of interactions of telephone support centers will be taken into load by conversational agents. We are 20% to date. (Source: Wikipedia / D J Shin)
However, we must believe in this fashion? “It’s been a long time, over 35 years, are known as algorithms that make reformulation” tempers Jean-David Benichou, founder and CEO of ViaDialog. “And in the contact center, the conversational agent on the web are used for 10 years quite effectively.” Indeed, recent years have seen flourish pop up on e-commerce sites, where an “adviser” is ready to lead you through a chat in your choice.
These robots are far remained to be
Indeed advisor is a robot that address and converse with you. Disappointed? Well aware that this is the same thing … phone. Apple’s phone support center uses the conversational agent to mark the first part of the customer experience and qualification needs.
These robots are far from being remained. “A lot of calls are treated this way,” says the founder of ViaDialog, actor Customer Interaction Management (CIM) since 2004. “And all the qualification and the formulation of the first response is provided by the conversational agent”. Precision work often because customer demands are complex. For the simplest applications, everyone will answer directly by tapping Google
Under the hood is the speech-technique to-text that is first used. The text then goes into an analysis engine. Conversational engine looks for the right answer, which is then transcribed via the text-to-speech technology. And this is no small matter. Because to seem “natural”, the conversational agent must be very reactive.
The reactivity at affordable prices
And it is this responsiveness, now at an affordable price which allows the proliferation of ads in this sector. “Since the necessary computing power was very important, the price of this type of system was unaffordable. With cloud computing, the operating cost becomes interesting is it that really changed, “said Jean-David Benichou.
Finally, these solutions are increasingly vitamin to Deep Learning. Technology that improves the natural conversation learning time machine to improve behavior to the customer. what develop new uses in an increasingly important context of mobility. the digital concierges are increasing and are “more relevant” recalls Jean-David Benichou. a relevance made possible by the use of information Contextual contained in smartphones, such as geolocation.
the call center operator still has a bright future
business side, it is obviously support centers and call center who can benefit from these innovations. “Today we can automate 20% of interactions,” says Jean-David Benichou. a proportion which applies to public issues such as banking and insurance, or customer subscription management to water or electricity. “And in 10 years, we are 50%” states CEO ViaDialog.
For computer support, however, where the corpus are not as standardized, it is still far from be the case. “We must lexical libraries with hundreds of words, not thousands of words” to be effective details Jean-David Benichou. Consequence: the call-center operator still has a bright future.