nib head of emerging technology Mathew Finch explained that nibby was envisaged as a way of improving live chat. The idea was to free customer service agents from having to handle so many simple inquiries, while ensuring that the chatbot could direct customers to an appropriate agent when human assistance was necessary.
Over its first year, nibby has handled about 21,000 chat sessions, completely handling 25% of them and routing the rest to the right area within the contact centre. This has saved some 500 hours of consultants' time, as well as delivering the information customers were seeking more quickly.
Developed with input from contact centre staff, nibby was deployed within a matter of weeks, Finch said. It was one of nib's first greenfields applications to go into the cloud (as opposed to existing applications that have been migrated to the cloud).
New applications are more easily built for the cloud, he observed.
nibby uses a number of AWS native services, including Lex (a deep learning chatbot service), Lambda (serverless compute) and Cognito (authentication, authorisation and user management).
"nib has taken an agile approach, using AWS machine learning to solve the business challenge of improving their customers' experience," said AWS ANZ managing director Paul Migliorini.
"Only one year into their journey with ML, nib is already proving that the AWS cloud can not only help them remediate technical debt, but is the ideal platform for genuine, customer-focused innovation."
While the timeline is uncertain, nib wants to move out of its data centres and into the cloud completely, Finch said, noting that all of the business's current systems could be migrated to the cloud.
Disclosure: The writer attended AWS re:Invent as a guest of the company.