By Masood Raza
Disrupting Internal Processes in Insurance
Several insurtech startups, such as Lemonade, Elafris, and others are reinventing existing processes to make things move fast and execute more accurately with various applications of AI.
These trend-setting companies are focused on delivering value through software that literally speaks to their customers and are acting on the data they are getting to make their experiences better overall.
Research by Willis Tower Watson determined that in the first quarter of 2019, insurtech companies received $728 million in funding, doubling the amount funded by investors in Q1 of 2018. A substantial portion of that is expected to go into developing AI-based solutions for better serving policy owners and stakeholders.
Underwriting automation is making leaps and bounds, with its ability to automate the process of identifying demand, new product creation, and accurate pricing and policy rating. However, relying on risk models alone can be a risky endeavour.
Whenever you are using a risk scoring system to reach a binary decision (e.g., approve or decline) there is the potential for things to go wrong. Underwriting automation systems for products with complex acceptance criteria are a work in progress and must be monitored by human underwriters and actuaries who build the models for accuracy and make sure they are making fair and acceptable decisions.
Putting Customers before Products
Most policy servicing and claims management processes are slow moving in nature and usually involve several manual steps until completion. AI can be put to work to increase the quality of claims assessment, create efficiencies in administration, and detect fraud.
An online lender, Upstart, can analyze vast amounts of consumer data and utilize machine learning algorithms to develop credit risk models that predict a consumer’s likelihood of defaulting on their loans. Their technology will also be licensed to banks for them to leverage in their underwriting processes.
Another startup, ZestFinance, had developed its Zest Automated Machine Learning (ZAML) platform specifically for credit underwriting. This platform utilizes machine learning to analyze tens of thousands of traditional and non-traditional variables that are used in the credit industry to rate borrowers. The platform is particularly useful to assign credit scores to those with limited credit histories, such as Millennials.
How much automation is too much, though? Robots cannot demonstrate concern, empathy, or connect with people on the compassionate level that is sometimes required of certain situations. Advisors would argue that anything short of a human delivering a newly issued life insurance policy or death benefit payment to a beneficiary would be inhumane and potentially hurt the insurer’s brand.
Working Together with the Bots
When pathologists from Harvard developed an AI system to discover breast cancer cells, the tech did well, with accurate readings 92 per cent of the time. But this was still four per cent less accurate than what human pathologists were producing, at 96 per cent. However, when the doctors used the tool in conjunction with their own processes, the numbers jumped to 99.5 per cent accuracy.
Thus, it’s unlikely the use of robotic process automation or artificial intelligence will completely replace human advisors anytime soon. That said, when insurers and advisors leverage the tools that automate certain mundane tasks, they can focus on helping their clients and unearthing the things that really matter to them, not just making black and white decisions.
Masood Raza is director of marketing with Crowdlinker, a digital product studio based in Toronto. He can be reached at email@example.com.