NAB's UBank gives face and voice to its home loan bot

By on
NAB's UBank gives face and voice to its home loan bot
This is Mia, UBank's new 'virtual human'

Third evolution of its bot journey.

NAB-owned UBank is expanding its use of bots to help customers complete home loan applications, creating a ‘virtual human’ called Mia that customers can talk to.

The bank will officially unveil Mia today before the service is showcased at IBM’s Think 2019 conference in San Francisco later this week.

UBank first created a bot called Robochat - powered by IBM Watson - to help customers get answers to home loan questions in real-time - in the hope those customers could then complete online applications faster.

Within 10 months, the bank was already touting “outstanding results”, with high usage, accuracy and - more importantly for UBank - a 15 percent increase in the number of applications being completed.

Now, the bank is set to evolve that customer experience even further.

It will still offer Robochat for text-based Q&A but from late this month, it will also allow some applicants to talk directly to a ‘virtual human’, asking questions in natural language and receiving answers in return.

The ‘virtual human’ is powered by technology from New Zealand company FaceMe, which itself uses some IBM components (though, the vendor says, the technology is capable of using other AI engines as well).

UBank’s experience with Robochat - and the fact it is based on Watson - meant that designing and implementing FaceMe has been reasonably streamlined.

“They had quite a firm idea of an NLP [natural language processing] system they wanted to build out whereas sometimes you’re starting with a blank page and people haven’t scripted the Q&A and don’t know the scope of the conversations [they want to have],” FaceMe’s director of marketing Jody Boshoff told iTnews.

“UBank had already thought about the use case, and because of Robochat they had a lot of questions and answers already scoped. It made it a quicker process.”

The Mia ‘virtual human’ will initially be available to 10 concurrent users while it is in a trial phase.

Both Mia and Robochat will be offered to customers as ways to get questions about home loan applications answered.

“There’s been a lot of testing and iterations of Mia to get ‘her’ to the point where everyone is confident in the experience,” Boshoff said.

“What [Mia] does is it takes Robochat but it makes [the experience] face-to-face.

“A home loan is one of the biggest and most stressful decisions you can make, and having that face-to-face engagement helps [UBank] to build emotional connection with the customer.”

Mia is initially designed to “answer more than 300 of the most common questions customers have about the home loan application journey”, the bank said in a statement.

“By bringing Mia to life, we’re giving customers a whole new way to interact with their online home loan application,” UBank CEO Lee Hatton said.

UBank has always said that its use of bots is not about reducing the number of actual customer service staff it employs.

That remains so, though Hatton said in a statement that the use of bots would mean extra staff were not required to meet any future rise in application numbers.

“We want to continue attracting customers but maintaining the same number of team members to support this ever-growing customer base,” Hatton said.

“That means we need to leverage key technologies like AI to tackle the typical questions customers ask, so we can free up our team to address the unique situations our customers need more support with, every day.”

In addition to Robochat, which was UBank’s first entry into the world of chatbots, the bank also has Robobrain, which acts as an enterprise search tool for customer service agents. Both are based on IBM Watson.

FaceMe provides similar digital assistants to other large brands in the region including New Zealand’s ASB Bank, Vodafone NZ and Auckland Airport.

Got a news tip for our journalists? Share it with us anonymously here.
Copyright © iTnews.com.au . All rights reserved.
Tags:

Most Read Articles

Log In

Username / Email:
Password:
  |  Forgot your password?