By Roy Urrico
Eltropy’s new interactive on-screen avatar, an intelligent virtual assistant named Kaitlin, had a smashing debut on the FinovateFall 2022 stage at the New York Marriott Marquis in the Big Apple’s theater district.
So popular was Kaitlin’s Broadway debut that Milpitas, Calif., based Eltropy, which enables financial institutions to digitally engage in multiple ways, had to set up an additional demo in the Finovate exhibition area to showcase the capabilities for what it calls “the industry’s first humanized AI (artificial intelligence) assistant.”
Following the Finovate demo session Eltropy’s Senior Vice President Murali Mahalingam and Chief Revenue Officer Dave Norton provided Finopotamus with an inside look at the avatar’s features.
A Need for Humanized Intelligent Virtual Assistants
Eltropy’s Mahalingam said Kaitlin empowers credit unions and other community financial institutions to deliver a humanized seamless experience while reducing contact center volume and handle times.
“The reaction from the Finovate audience clearly demonstrates that our industry wants a humanized AI solution as the next evolution in bot technology” stated Mahalingam, the former vice president and founder of Marsview, which Eltropy acquired in August, 2022. “Our vision of delivering a platform that can incorporate texting, video banking, voice, secure chat, and cobrowsing enhanced by AI, sets us apart from the industry and delivers a personalized experience for consumers. Eltropy acquired Marsview specifically to utilize AI and create the industry leading platform for humanized AI.”
Explained Mahalingam, “All these banks and credit union have websites where they put in a lot of investment in content that exposes all their products and services. But the only way a member or a customer can interact with (financial institutions) is through a chat channel from the website. We are changing that today by adding a digital voice capability and an avatar that can really create that human connection.”
Eltropy maintained while consumers have become accustomed to interacting with a chatbot or voicebot, it is often too robotic and impersonal. This is especially the case for credit unions that pride themselves on delivering exceptional service with a personal touch. According to Mahalingam its new AI solution allows members to have natural and efficient conversations, almost like speaking directly with a service representative.
Kaitlin can listen and talk like a human, removing the need for keywords or long menus. With advanced natural language processing and machine learning, the virtual assistant gains intelligence with each interaction, and can be customized based on each financial institution’s needs.
Kaitlin’s Appeal to Financial Institutions
Mahalingam explained in general session demo that Kaitlin kept up with the conversation’s progression and back-and-forth interchange. The inbound caller (Mahalingam) expressed his first objective: “I wanna make a loan payment.” And then he shifted the context from making a payment to requesting an account balance. The avatar retrieved the balance and then remembered the previous context. Hey, you wanted to make a loan payment. Do you want me to go ahead and help you with that?’ That is the power of a multi-intent recognition model that we have built, where it can understand multiple contexts and recall context for you,”Kaitlin responded.
Said Norton, “Most bots go down a sequential line and then all the way back; she was jumping from one to the other.” Norton added, the avatar can remember current details, or a conversation from a month or even a year ago. “No other bot can to do that.”
Norton suggested Kaitlin should appeal to any financial institution that has digital customers seeking a more personalized, more cultivated virtual interaction, “because we crave humanlike connection.” He offered internet robots can only automate so much. “That is where this avatar is very intelligent, one of a kind, can recall information, can have empathy, and compassion and ask ‘how are you doing?’ and things like this, in a very natural conversation.”
Kaitlin also the capability to verify identities on the fly. Mahalingam said, “There were two (authentication) factors that were in place (during the demo). It asked me to enter the phone number. I entered the phone number. So, she found out who I am, but she still had to verify is that the same person? Then we use voice authentication, voice biometrics. She said, ‘Read me this phrase,’ I read out that phrase. So, it recognized me as Murali [his first name] based on the phone number and then validated that it is Murali who is talking to me.”
During the demo, Norton inquired how many businesses ask customers for their name, password, user ID, social security number and then do not carry that basic information in the next step. “You are always having to restate it over and over. It gets you so frustrated as a customer. How many times do I to tell my same story, whatever it might be.”
In the Finovate demo Kaitlin used voice recognition for authentication but facial recognition is on the table for other Eltropy AI solutions, Mahalingam confirmed. “We have another AI, which we did not demonstrate today that can do face recognition. It can see the person and say, ‘Hey, I'm talking to Murali.’”
While the curtain has risen for Kaitlin, it is slowly auditioning for a starring role within Eltropy’s client base, which includes almost 500 financial institutions, 400 of which are credit unions. Norton pointed out, “We are just launching it and we'll be inviting or accepting different institutions that want to be beta customers.”
Mahalingam also explained Eltropy is taking a gradual approach to launching Kaitlin. “This is something we will work on a case-by-case basis. The product from a technology perspective is ready, but it must be relevant for the business, the financial institutions,” said Mahalingam. “So, there is some work to be done there.”
Norton noted there is more that take places backstage to set up Kaitlin. “We have to work with (financial institutions) to create their vernacular, their terminology.”
Right now, the avatar is trained to answer questions, complete business transactions, and perform tasks like setting up an appointment or making payments. However, Eltropy intends to create different use cases for financial institutions for both inbound and outbound interactions.