Australia’s two largest group insurers plug into AI

By

14/11/2017

Insurers expect artificial intelligence to enhance, rather than replace, their workforces, as they trial technologies for operations.

AIA Australia is piloting Watson’s chatbot and TAL Life has an in-house team developing AI solutions with a view to improving the customer experience, executives told the 2017 Conexus Financial Group Insurance Summit, held in Sydney on August 29.

The Watson cognitive computing platform, which mimics human decision-making, but more quickly and using more data, is being applied by 20 industries in 45 countries. Insurance is in the early stages of using the technology, delegates heard.

But both companies say AI is more likely to enhance than replace roles within their companies.

AIA group operations head Derek Vine said a recent pilot of chatbots initially caused apprehension among some members of the claims team, particularly older staff.

“When we did our first pilot, I sat in a room with 80 claims staff,” Vine recalls. “I’d say 60 to 70 per cent were Gen Ys, like myself, and were very excited. Some were less excited about that change management and fear of jobs being lost.”

Vine is adamant that rather than replacing the human workforce, the efficiencies robots will create are more likely to set people free to perform more meaningful work.

“We are nothing without our people. You’ll never be able to replace a conversation, but AI will enable better conversations,” he said.

AIA, one of Australia’s largest life insurers, has trialled Watson performing medical claims assessment and said that in two weeks the AI platform had learned to ingest policies and make accurate decisions with a degree of efficiency it typically takes a human assessor up to five years to achieve.

“As soon as Watson was able to learn all that, it had 99 per cent accuracy against the assessors’ decision,” Vine said.

He said Watson had the ability to create “smart rule sets” that are driving productivity in the assessor space.

“We started with a mantra to improve customer-centricity and support claims staff [in being] the best they can be,” Vine said. “AI doesn’t solve everything, but it helps.”

AI in back office and front office

AIA’s second pilot with Watson is for making it the financial calculations on claims inputs, such as pay slips, offsets and partials – calculations that have a relatively high error rate when staff make them. Those trials are still in the early stages.

“If we can automate that, we’ll improve productivity,” Vine said.

Chatbots with empathy

TAL Life general manager, innovation, Dan Taylor, said an in-house team using another platform had developed two chatbots: Cora, dedicated to the claims journey, and Alfred, a life-insurance broker with a penchant for dad jokes.

Taylor said the team “thought about it in two ways: on the front end, how chatbots could improve customer experience, and on back end, applying machine learning and processes to help customer experience.

Cora was empathetic and offered tips and stories from former claimants, while Alfred operated around the clock, helping people get a quote and search for the right policy before setting up appointments with real people.

“There’s been a lot of hype on the potential for this to change the industry and take jobs,” Taylor said. “Actually, I think we’re a long way off that. While it’s good, it’s not 100 per cent accurate because there’s always a case where there’s not enough data.

“Using machine learning to predict underwriting outcomes is driving consistency into our underwriting process but I would advise [companies] to think very carefully about how [machine learning] can support processes, rather than [using it to invent them] from scratch. This machine learning vision of the future is still a way off.”

Insurance lawyers embrace AI

Fergal Murphy, from the IBM Watson industry platforms division, said insurance was “on the cusp of disruption”, but predicted AI would also be used in conjunction with humans engaging with one another.

“The best use of these types of capabilities is partnering with people. There are certain things we don’t do – intuition, morals and ethics,” he said.

He added that the team at IBM prefer to think of AI as “augmented”, rather than artificial, intelligence.

Insurers would need to create new roles and upskill staff, rather than replace their workforce, to incorporate AI and take advantage of its capabilities, Murphy said. However, he noted that disruption in the insurance litigation space was moving at a greater pace, with much paralegal work already automated.

The 2017 Conexus Financial Group Insurance Summit was produced with thanks to support from platinum sponsors AIA Australia and TAL Australia, and silver sponsor Swiss Re.