Getting Real about Artificial Intelligence in Insurance (Expert Q&A)
Last month’s announcement that Fukoku Mutual Life Insurance in Japan has fired employees and is replacing them with an artificial intelligence (AI) system that can calculate insurance payouts caused some concern in the insurance industry. Should insurance personnel be worried? And how can insurers best harness this growing trend?
To answer these questions and more, Sapiens gathered five of its leading subject matter experts (SMEs) for a quick Q&A to get real about artificial intelligence (AI).
What are some of the potential benefits from the applications of artificial intelligence for insurers?
Alex: Underwriting will improve tremendously. The machines will be smart enough to automate the underwriting process and issue policies quickly and accurately.
Igor: I agree and I think that AI can also be used in claims processing to detect insurance fraud. Using sentiment analysis, the machine will be able to detect fraudulent tendencies automatically. Of course, this will be backed up by powerful algorithms that will collect data from social networks and cross-reference with other systems in seconds.
Michael: The sales department can greatly benefit. Machine learning will help us pick up on customer preferences, so sales teams can tailor their products in a timely fashion.
Irena: To add to what Michael said, I think customer retention and the contract renewal, cross-sell and up-sell processes will become more efficient, as sentiment analysis becomes more widespread.
Sounds like a very happy, automated future. What challenges do you foresee?
Michael: Cyber-security, since we are dealing with the impersonal aggregation of customer information. This information will need to be secure, whether stored in the cloud, as a hosted service, or on premise.
Alex: Clearly, a new set of regulations will have to provide the guidelines for all players in this ecosystem. Where, how and when can AI services and results be used? Regulation is currently far behind where it needs to be.
Gil: The machines will have their elaborate algorithms to cross-reference fraudulent activity, calculate risk at the speed of light and signal an insurance agent if
a customer is close to disengaging. But we’re forgetting the fact that machines, like humans, can err. Considering the stakes, errors could be disastrous.
Irena: One of the current challenges is that AI is very expensive, so SMBs can’t currently afford it. But the technology is in its early stages.
Let’s cut to the chase: could insurance personnel be replaced by machines?
Gil: This is not on the immediate horizon for most insurers. I think it will free underwriters, actuaries, salespeople and others from simple computing and analysis tasks, and they will be able to offer a more personalized, holistic approach.
Igor: Machines cannot (yet?) effectively relate to humans, and that is crucial in insurance. Insurance clients turn to their insurance providers in sensitive situations, like major property loss, or deaths. I don’t think clients are ready to give up on sympathy and the human touch.
Michael: I don’t think the change at this stage will be about replacing humans, but rather pushing them to be better. The focus should be about harnessing the machine capabilities for an elevated customer experience.
Irena: I think this move will actually create jobs. IBM CEO Ginni Rometty spoke at a panel discussion in Davos, Switzerland this month and used the term “new collar” to describe jobs that will be created by this “Fourth Industrial Revolution.”
Thank you all so much for your input. Final question: give me your two cents on how insurers can prepare for a more “intelligent” future?
Igor: I think the best way for insurers to prepare is to organize their data and make it accessible. Today’s insurers are coping with mountains of data. They will need to quickly transform this data into business decisions by using advanced analytics tools, and this organized data will be the foundation that machine learning will build upon.
Gil: Insurers should start researching which technological partners fit their criteria. An experienced and proven partner can act as a trusted advisor to insurers, bringing best practices that have saved scores of other leading insurers time and money.
Alex: If they haven’t already, insurers should make the leap to pre-integrated, open digital networks that can talk to each other, eliminating integration efforts between the front- and back-end, and facilitating an end-to-end experience to simplify the customer engagement process.
Michael: First, ensure that data is protected. Secondly, management and employees must become as flexible and agile as the systems they will eventually work with.