NAB eyes Misty for 'tech lounge' robot

By

Coincides with expansion of employee chatbot beyond IT support.

NAB is proposing to do a proof-of-concept with a robot called Misty, initially to see if it would make a good first point of contact for employees dropping by one of the bank’s eight ‘tech lounges’ nationally.

NAB eyes Misty for 'tech lounge' robot

The bank runs the ‘tech lounges’ in office locations around Australia as a place where any of its tens of thousands of employees can get in-person technology support during business hours.

Head of technology for workplace support services, Raj Ghuliani, told a Microsoft Ignite session last week that the bank hoped to use the lounges as a venue to test the Misty robot.

Misty - made by a robotics company of the same name - made an appearance at Microsoft’s Build conference back in 2019, and has since appeared in demonstrations that also use Microsoft cognitive services. It is now owned by a Swedish company, Furhat Robotics.

Ghuliani mentioned NAB’s plans for Misty only in passing.

“We are trying to do a proof-of-concept via a physical bot called Misty,” he said.

“We’d like to see if we can utilise that for triaging purposes, at a start, for our face-to-face tech support, what we call our tech lounges.”

Most of Ghuliani’s presentation at Ignite was focused more on a virtual agent the bank has been developing since last year.

It is currently called ‘NAB Bot’, though Ghuliani said the bank planned to let staff rename it once its target state is achieved.

NAB Bot’s initial use is as a tech support chatbot that is trained to respond to 800 common questions or requests posed to the bank’s tech support teams.

Ghuliani said the bot had so far handled about 57,000 requests that would otherwise have resulted in calls to the tech support desk.

The bot, which is built on Azure Cognitive Services and Bot Framework, is integrated with ServiceNow, which powers NAB’s knowledge base and live chat functions.

Ghuliani said NAB Bot can direct employees to knowledge base articles, or - upon request - transfer an employee and their chat history to live chat for extra assistance.

At present, employees interact with NAB Bot via text-based queries; it is trained to recognise 18,533 “unique utterances” - phrases typed in by the employee. 

“We are also working on implementing voice support via the NAB Bot,” Ghuliani said.

“Our current bot works via being able to type as part of natural language understanding and also by being able to click on the options and select from the menu. 

“Moving forward, colleagues will be able to talk to the NAB Bot and give NAB Bot instructions or ask questions openly. This gives our colleagues a seamless experience across all support channels.”

NAB Bot is also presently set up as an additional tech support channel, in addition to all existing ones such as phone-based and live chat support. 

Employees make the choice on which channel they use.

The target state for NAB Bot is that it becomes a first point of contact within NAB for all “work-related queries”. 

This would see it expanded to include human resources, property services, procurement, business apps and compliance FAQs.

“All this information is available at many other locations, whether on the intranet, or other policy pages, but for our colleagues to be able to find those it can take them a few minutes,” Ghuliani said.

“We’re trying to save them that time by letting NAB Bot give [them] the right answer.”

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

Most Read Articles

RBA reveals three-year project to upgrade payment IT systems

RBA reveals three-year project to upgrade payment IT systems

CBA backs GitHub automations to get new features to customers faster

CBA backs GitHub automations to get new features to customers faster

NAB decommissions 26-year-old Teradata platform

NAB decommissions 26-year-old Teradata platform

BoQ pressured to reveal automation impact

BoQ pressured to reveal automation impact

Log In

  |  Forgot your password?