Suncorp has introduced IBM Watson into its online claims system that processes more than 500,000 motor claims every year, to help the insurer better understand the circumstances of the claim and determine liability.
Using IBM Watson's Natural Language Classifier, the system analyses customer descriptions of motor vehicle accidents; however, as the descriptions are often written in a conversational way, IBM said they often include colloquialisms and Australian slang that Watson will need to learn in order to properly assist Suncorp.
The technology is being used across Suncorp's Personal Insurance brands comprising AAMI, Suncorp, GIO, and Bingle.
The platform relies on Watson to conduct liability analysis and assist in fast-tracking simple claims, such as single vehicle incidents with detailed descriptions. However, if the incident description provided by the customer is limited, or there is a low confidence score from the system in determining liability, the claim is routed to a human to help with the decision.
The AI tech began production in June 2017, and since then the proportion of customers who were fast-tracked through the process has tripled, said CEO of Insurance at Suncorp Gary Dransfield, with Watson enabling claims to be lodged, excess paid or waived, and repairs booked within five minutes.
"This technology augments our claim consultant's knowledge and expertise, providing data-driven insights and instilling greater confidence in our liability decisions. It has also helped us speed the process for our customers, while improving their experience at a time that's often very stressful," Dransfield said in a statement on Wednesday.
"Ultimately this enhances our customers' experience by delivering innovative digital solutions to help make the claim process easier."
According to Dransfield, relying on technology to make the decisions will ensure the claims process is streamlined. It will also result in a more consistent approach to liability decisions, as Suncorp hopes Watson will establish a reliable reference point based on historical claims and industry guidelines.
Suncorp trained Watson on the complexities associated with determining liability in motor vehicle accidents using nearly 15,000 de-personalised claim scenarios and the resulting liability determination, in line with guidelines from the Australian Prudential Regulatory Authority (APRA). By the end of the trial period, Watson could accurately determine liability for around 90 percent of cases, IBM said.
In delivering its results for the 2017 financial year, Suncorp CEO and managing director Michael Cameron told shareholders the group's AU$394 million after-tax profit would have been higher had it not needed to pump further funds into its core banking platform, contracted to Oracle.
The bank previously had its eyes set on reaping AU$170 million in benefits by the 2018 financial year from its technology optimisation program after announcing back in May 2014 that it would be dedicating an estimated AU$270 million to changing its core banking system.
Cameron explained Suncorp had decided to leave a small group of its products running on the old system, rather than take a risk in launching a new platform.
"At the end of the day, the customer experience is what counts; we don't want to be in a situation where we're getting outages so all we're really doing is waiting for a new version of the same software to be released, which incorporates the sorts of things that we want, rather than build them ourselves internally," he said.
"It will actually cost us a little less, it will de-risk the operations from a customer perspective, and for us it means running two systems for a period of time. While it's not ideal, it's something that we've thought long and hard about."
PREVIOUS AND RELATED COVERAGE
Pip Marlow joins the bank as its new innovation-focused chief executive officer.
Heading down a data science path kicked off an innovative culture within Woodside Petroleum, but most importantly it helped solve some business problems the exploration giant was struggling with, its CDO has said.
IBM's Watson is impressive technology, even if it can't quite cure cancer. It's time the company positioned it accordingly.
IBM's plan is to create a data science operating system that can bring together data scientists, analysts, and business leaders.
As part of an expanded partnership, IBM will participate in Docker's Modernize Traditional Applications, and Docker Enterprise Edition will get an integration for IBM Cloud.