Embracing a new way to see data
Taking a steady and intentional approach to adoption can help insurers successfully implement AI and machine learning.
A survey from IBM shows that 41% of respondents switched their insurer in the previous two years due to a slow reaction to their changing needs.
And yet, despite these departing customers, insurers have remained slow to embrace emerging tech like artificial intelligence (AI) and machine learning (ML). These tools cannot only help them better serve their users but also can allow them to operate more efficiently internally. With the wealth of data insurers have at their disposal, bringing in cognitive computing tools to make it actionable is an essential move.
However, in order for the insurance industry to address their inefficiencies using AI and ML, its leaders must evolve their outlook.
What’s holding insurers back from emerging tech?
Insurers are set up particularly well for using ML and AI because of the sheer amount of data and research they must retain. They’ve built up their internal knowledge and datasets for so many years that they’ve already scaled one of the top barriers of working with AI — having the data volume to drive learning and insight.
However, the insurance industry takes a traditionalist approach when it comes to tech. The sector’s conservative leaders have worked to put many tools and sophisticated statistical models in place. From their view, if the system is generally working as is, why would they pull the rug out from under it to overhaul it? Doing so could be risky — something insurance organizations are trained to identify and manage.
Of course, they should plan for — rather than resist — these kinds of changes in the name of digital transformation. I believe things are moving in the right direction slowly but surely.
How insurers can move toward AI’s benefits
Getting started with new and powerful tech like AI can be intimidating, especially for an organization with a traditionally reluctant approach to new tech adoption. However, taking a steady and intentional approach to adoption can help. Here are some fundamental steps leaders should follow as they move to integrate AI into their processes.
- Understand if your problem is relevant to machine learning’s capabilities — Machine learning and AI hold a lot of potential for the insurance industry, but they are not a catch-all solution. Thus, an insurer’s first step should be to consider the problems they’d look to apply it to. If data is a main consideration in the problem — such as if it concerns payout values or claims processing — then machine learning likely has a logical application since it requires huge datasets in order to serve up insights. However, not all pain points can be addressed by ML, and imposing the technology where it won’t drive measurable results can be a costly and time-consuming mistake.
- Decide what you want the solution to the problem to be — When you identify a pain point you believe AI can assist with, you have to decide how you want to change the outcome, and to what degree. Understanding what you want the outcome to be affects how you’ll teach and manipulate the machine learning. For example, if your payout values have skyrocketed in the past year, you should know if you want to return the values to their previous rate, or if you’d like to lower them further to affect the organization’s budget.
- Partner up with an expert — As an emerging technology, machine learning requires specialist knowledge. It’s not recommended that a team spend a significant budget to DIY a machine learning solution, only to come up with subpar results. Machine learning tools often demand more specialized expertise than most in-house developers have. Therefore, a company should consider linking up with an external team of experts who have plenty of AI experience. Understand it’s an investment — if you do it, be fully committed.
While insurance may be one of the more traditional industries, that’s changing as emerging tech like ML and AI becomes more democratized across the business to business world. If insurance can scale the traditionalist attitudes throughout the industry, its affinity for data makes it a perfect candidate for becoming a leader in using AI to solve problems.
Ray Johnson is the chief data scientist at SPR. Contact him at Ray.Johnson@spr.com.
Related: