Conversational AI improves customer experiences by reducing friction and providing high-touch service at scale.
In the summer of 2020, IDC surveyed more than 500 end-user organizations, and found that “for contactless experiences, voice-based interfaces were the top planned investment” among respondents.
Voice-based technologies have made significant improvements since the days of systems such as Moviefone, said Hayley Sutherland, Senior Research Analyst at IDC during the latest Amelia webinar, “Conversational AI and Friction-Less Customer Experiences”.
Sutherland said that IDC has seen the “improved accuracy of these language-focused AI systems.” These systems have become increasingly better at recognizing different accents and “different linguistic challenges, like idioms.” As voice-based systems like speech-to-text and natural language processing improve, interactions with AI become more human-like.
“These voice-based AI systems are getting better at not just understanding the actual utterances, you know, ‘what are the words that people are saying,’ but also user intent, context, and sentiment,” Sutherland said during the webinar. Humans are complex, and although we sometimes seek simple yes or no answers, we often require a full conversation with an agent in order to unpack all of our questions and concerns.
To make sure your voice-based interfaces contribute to frictionless experiences within your organization, Sutherland highlighted a few key points to keep in mind when creating a voice-based conversational experience. To ensure users don’t hang up or ask for a human agent, “you need accurate speech technology,” and must aim to “achieve that personal touch at scale.” To ensure customers can still speak to a human, your AI system should be able to loop in human agents if they are needed or requested.
Sterling National Bank and Skye
Sterling National Bank is a relationship-driven bank. As Jason Vazquez, EVP and CIO of Sterling National Bank, said during the webinar, the bank does not bounce customers around when they need help. The goal is to assist its customers with a “single point of contact,” however, up until this past year, the bank was still using IVRs (interactive voice response systems) and contact center agents to handle first-touch customer calls. Although contact center agents were delivering high quality customer experiences, customers still needed to spend time in queue, which can impact customer satisfaction.
When Sterling was determining its vision for deploying Amelia, renamed Skye, they began by thinking about the current customer friction-points that could be addressed with Conversational AI. While the bank initially wanted to place Skye between their IVR and their contact center agents, Vazquez said the bank later realized that “the more frictionless way to handle her was to present Skye to every single attempted call, and let her resolve most of that conversation.”
The bank’s approach with its IVR was to deflect calls to an agent, but they changed tactics with Skye. Vazquez said the Conversational AI is “meant to enhance” the experience to ensure things are “fully resolved,” and since Skye’s deployment, she has delivered on these expectations.
For example, Vazquez said that many people call the call center to determine their bank balance. “The irony of that is that they’re typically using a mobile phone to call for their balance. On that same mobile phone, they would have the balance on it from their mobile app,” said Vazquez. “Are they really calling to check their balance, or is there something deeper of their intention to call?”
While IVR systems may hang up after providing customers’ bank balances, Conversational AI systems converse with customers end-to-end to ensure all of their questions are answered in one call. In doing so, Conversational AI systems reduce friction points by making sure that customers don’t have to keep calling back.
Vazquez shared that Sterling has passed a significant milestone with Skye: She has handled more than one million conversations. In addition, “Skye is resolving more than three times the amount of calls that contact center agents can handle,” said Vazquez.
The Days of Simple FAQ Chatbots Are Far Behind Us
Joshua Schechter, Amelia's Director of Cognitive Implementation, provided a variety of practices during the webinar to help guide companies through a successful Conversational AI implementation. His first piece of advice was to start with a vision, as Sterling National Bank did, and consider questions such as: Why are you choosing Conversational AI, who will be using the technology, what are you building and when will you deploy it?
“We are way past the days of building bots to answer simple questions,” Schechter explained. “Let’s actually build something that can take action, and that’s helpful, and resolves user issues.”