Financial services companies are investing in voice and chat conversational technologies — but can they really drive engagement and streamline customer communications? Catch up on this VB Live event to learn how Capital One and others are leveraging these technologies to create meaningful customer experiences.
Consumer enthusiasm around digital assistants like Amazon Echo and Google Home sparked a run on investment in the financial services world. Financial institutions have been going all in on conversational technologies such as voice and chat in banking.
“There’s the near-term focus that a lot of folks gravitate to: How can we rip costs out of the system by having people interact with automation and AI instead of a human?” says Ken Dodelin, VP conversational AI products at Capital One. “There’s another side of it, too. As we make advances in AI and we’re able to predict what customers need, we can get to them proactively.”
They’ve also found that there are certain types of transactions that can just be much more productive with a voice- and chat-enabled technology than with a human being. For example, a bank wire, where the human being can’t check the ABA routing number as fast as a computer can.
“We’re counting on the productivity of the digital interface also being a preference for consumers and businesspeople,” he adds.
But why hasn’t the financial services industry been completely transformed yet? Answers run the gamut, from corporate politics to where and how to deploy the technology, and security issues in a sector where the data being handled is highly sensitive. And part of it is simply the newness of the technology, and customer expectations.
“Due to popular culture, often we have customers coming on and using these conversational AI systems and very much expecting them to be of a human level of intelligence,” says Sandi Boga, director of product innovation at ATB Financial. “But we’re just not there with the state of the industry.”
So when the system gives the wrong answer, or a consumer experiences failure, they have a tendency to drop the interaction very quickly and become disillusioned.
“If we lose people at these early stages, it’s going to be very difficult to get them back over time,” Boga adds. “With a new technology that isn’t working as you expect it, you need to do a lot to get customers to get on board, and then if they get on board and experience failure, it’s difficult to bring them back later.”
In the early days of pioneering these technologies, companies need to find a balance, he says, between growing in sophistication while still bringing customers along with you on that journey.
“The whole idea of designing the conversational user interface — it’s quite a new concept to some of our UX and UI people,” Boga says. “We’re learning as we go, learning from humans, and also learning from the capabilities of the AI.”
Dodelin agrees, pointing out their recent insights around their bot-powered SMS fraud alert.
“There’s a big behavior consumer change that needs to happen here with this new technology,” he says. “Yet there are some cases where it’s actually the complete opposite. I like to say, we launched the chatbot and nobody told us about that.”
When the fraud alert goes out, customers were supposed to respond with confirm or deny — the bot doesn’t understand any other answer. But when they looked at the logs, we found, not surprisingly, that people don’t behave like machines. Fifteen percent of replies weren’t just confirm or deny. They were seeing replies like, “yes,” “no,” “yeah, that was me,” “that’s baseball equipment I bought for my grandson,” emojis, and more.
Adding in natural language processing bumped the chatbot’s understanding of the variety of possible replies up to about 99 percent, he said.
“People were already using the conversational interface in text, and now we’ve just adapted the technology to help the machines understand,” Dodelin explains.
“Being able to add on additional clarifying questions and asking the person for a bit more information in order to give an answer that’s more expected is a really interesting evolution,” Boga adds. “We’re moving away from replicating a Q&A experience on a chatbot interface and toward actually designing a new interface to be used on these computers. The evolution is continuing to move forward, which is very exciting.”
“[With] the complexity and personalization that voice and chat will have a year from now, I think we’ll get close to the point where you’ll see a tipping point, where human beings will like the interaction,” Frank Coates, executive managing director at Envestnet | Yodlee Analytics says. “They’ll start to understand that a high-quality, personalized discussion with voice or a chatbot is going to be very productive and very humanistic.”
And companies who have resisted the early-adoption stage are starting to feel the tipping point coming on as well.
“I feel like we’ve reached the top of the hype curve and it’s starting to come down a bit, but everyone’s starting to take notice of the implementations that have already been done, with folks like Capital One, USAA, and ATB Financial,” Boga says. “They’re starting to see the applications for their own businesses.”
To learn more about how AI-powered conversational technologies can capture the hearts of financial customers, the psychological advantages of chatbots, and how to launch a successful conversational technology strategy now to ride the coming zeitgeist, catch up now on this VB Live event!
Don’t miss out!
Webinar attendees will learn:
- How conversational technologies changing the financial services landscape
- Things to consider when looking to deploy a voice-enabled solution
- How voice and chat can support a broad range of financial wellness solutions
- How to capitalize on conversational interfaces to capture and improve customer loyalty
- Frank Coates, Executive Managing Director, Envestnet | Yodlee Analytics
- Ken Dodelin, VP Conversational AI Products, Capital One
- Sandi Boga, Director, Product Innovation, ATB Financial
- Evan Schuman, Moderator, VentureBeat
Sponsored by Envestnet | Yodlee