At this year’s Digital Workforce Summit, Verbio co-founders Jordi Torres and Carlos Puigjaner led a breakout session on the current state of conversational AI technologies and how they will develop in the years to come.
At this year’s Digital Workforce Summit (DWS), Telefonica’s Global Contact Center Director delivered a keynote presentation on the main stage in which he summarized his company’s automation journey with its Peruvian contact center via Amelia. While Amelia provided the cognitive underpinnings and back-end integrations for the implementation, the project could not have succeeded without Spanish voice technology specialists at Verbio who helped build a seamless conversational bridge to the call center’s Spanish-speaking clientele.
Verbio co-founders Jordi Torres and Carlos Puigjaner led a breakout session at DWS on the many ways that conversational Artificial Intelligence technologies like Amelia are transforming the way business is done across multiple industries and languages.
“Typing will always be there, but voice is the new driver of interaction between machines and humans,” explained Torres. “Because voice makes things easy, convenient [and] on the go. You don't need to use your hands; you don't need to use your fingers. You're just cooking and using your voice, driving and using your voice. Voice will be part of any aspect of the hyper-connected consumer.”
Drivers of Conversational AI Adoption
Conversational solutions benefit consumers by providing hands-free and seamless interactions, while subsequently automating routine processes for an enterprise that utilizes such technologies. “Our customers have been looking for two different things,” said Puigjaner. “First is improving the customer experience, and the second one is operational efficiency – that has to do a lot with cost reductions.”
While reducing overhead costs is a substantial benefit of conversational AI, it’s by far not the only one. Since consumer engagements take place via a digital medium, it opens the door to advanced functionality that a human agent cannot provide – such as biometrics, which can be used to independently authenticate or identify individual users, Puigjaner said. This kind of functionality has obvious utility as a layer of security, but it can also provide additional functionality to consumers.
“Imagine arriving at your home with Amazon Alexa and saying, ‘Put on my playlist,’ and Alexa will understand that it's me asking and it's not my wife or my daughter,” Puigjaner explained. “That can be powered with voice biometrics.”
Voice as Fraud Prevention
While many authentications can be handled with the use of passphrases, a more advanced biometric feature that Verbio is working with is “free speech,” in which the system automatically identifies the speaker just by looking for unique audio fingerprints. Free speech has certain challenges – notably that it can take up to 30 seconds to authenticate, compared to a simple passphrase which can take as little as four seconds. Systems using free speech must also be able to identify fraudulent inputs from a recording of a person’s voice. Research has shown that humans never repeat the same phrase exactly the same, which provides an opportunity for these systems to root out potential frauds, with the right software, he said.
“So, if we detect that [a phrase] is something that was recorded [exactly] before, [the system] will deny your entrance, or will ask you again to repeat it again,” said Puigjaner. However, free speech identification also provides many advantages such as the ability to create a public audio “blacklist” of potential fraudsters shared among companies. If systems recognize the voice of any individuals on these blacklists, they will know to keep them out of their systems.
Conversation is, of course, a two-way street. These technologies must be able to understand user inputs and also provide digital systems a voice of their own. With such considerations in mind, companies must also think about the branding potential these technologies provide.
“AT&T doesn't want to have the same voice that you have when you talk to Sprint, T-Mobile or Verizon,” said Puigjaner. “Every brand wants to have their own voice; they don't want to have the same voice… For the customers who say, ‘I want my voice,’ we start by asking, ‘Okay what do you want your voice to be like? You want your voice to be spontaneous? You want them to have expressivity? Or you want it to be super professional?’” Some customers want all of that, and then also want it in multiple languages, “but we are able to do that,” he said.
Cognitive AI provides a means for humans to engage naturally with digital systems, and pioneering companies like Verbio and IPsoft are leading the way to making these engagements as useful and as valuable as possible.