Toronto, ON (Dec. 17, 2017) – Artificial Intelligence (AI) is getting serious attention by insurance practitioners. There is an expectation that AI will bring new opportunities for market segmentation, underwriting, risk management, claims, and administration. But the scope is large and can cause paralysis by analysis. We’ll look at two examples … one that is just starting and one that has matured.
From the Analysts’ view
Mark Breading, Partner at SMA, produced a report – AI in P&C Insurance. Breading noted that AI has become the hottest topic in insurance technology. The immediate result, according to SMA, is that “More insurers plan to invest in AI over the next five years than any other emerging technology – with 80% planning to spend on the technology.”
The SMA report breaks AI down into specific functions and their utilization. For personal lines, the top three functionalities are:
- Data Text Mining (67% of insurers surveyed)
- Machine Learning (52%)
- Chatbots (42%)
Commercial lines has one slight difference:
- Data Text Mining (77%)
- Machine Learning (66%)
- Robotic Process Automation (51%)
Breading notes that Data Text Mining – the leader – offers significant potential in analyzing unstructured text and offering new insights on customers, risks and operations.
Breading also describes relatively quick benefits from Machine Learning because it can be used “both as a standalone technology for some applications basis as well as an enabler for other technologies in the AI family.”
In other words, these are important structures in the world of insurance AI.
Where does value come from?
From a functional standpoint, the SMA research found that the two top business cases were Underwriting Decision making and Customer Experience. However, Underwriting was highest in commercial lines while Customer Experience was highest for personal lines.
The next two highest functions were Fraud Detection and Claims Adjusting, These came in the same order for both lines of business
Looking five years forward….
Insurers told SMA that they believed that underwriting and claims will be impacted the most over the next 5 years. SMA explained that this was due to the centricity of these functions in the business of insurance and the reliance on human experts.
Going forward, SMA recommends that AI “should be included in every insurer’s strategies and plans.” From an operational perspective, SMA recommends that “The best approach is to find a strong alignment to your business strategy and identify specific uses of the technology that will add value for your company.”
Some Financial organizations already have the five years
Schwab, the discount broker, has been active implementing AI for some period of time. In a recent interview with Adam Lachinsky, executive editor for Fortune Magazine, Walt Bettinger, CEO of Charles Schwab shared his views of the AI experience to date in the company to date.
Schwab is a large, unique organization. Bettinger notes that, in spite of its focus on a being a no-tradeoffs, discount firm, “J.D. Power ranked us No. 1 in full service for brokerage. Who would have thought 40 years ago Charles Schwab would be No. 1 in full-service brokerage?”
How does technology fit?
In short: With scale and time. Bettinger says, “We have a long term time horizon”. Within that view, Schwab technology has a unique role. According to Bettinger,
When you’re responsible for custody of $3.2 trillion, technology has to be at the forefront of virtually everything you do. And not just for efficiency, but also for minimization of errors and processing, for protecting clients’ information, security of their personal information and their assets. Then you have to back it all up with a guarantee.
And AI plays a unique role….
Bettinger looks to his strategy to make optimum use of AI, and finds that AI works best in a service capacity. He says, ” We have a saying at Schwab that we use our data to serve, not to sell.”
Bettinger provides an example:
You call up one of our call centers, and you have a concentrated position in XYZ stock because you worked there for 35 years. Our computers can listen to that call, interpret what you are saying, and reach right back out to you with an option strategy.
So what did we learn today?
While insurers are determining opportunities for AI, there may be good value in challenging existing roles and goals. With that under the microscope, there are more precise opportunities to define use of AI over longer term with more precision.
Is this possible in the insurance community? Are you participating in an AI initiative?