Lessons Culled From Early Innings of AI

AI is transforming claims, but there are pitfalls. Success requires extreme discipline about identifying and addressing a single problem.

The workers’ compensation claims process tends to creep along, often inviting lawyers to the party and leaving both workers and companies frustrated. But this trend is reversing. Why? Technology has been developed that alleviates the stress points that threatened to eventually break the workers’ comp system. The same is true for other insurance segments, as well. The technologies sparking this massive transformation across the insurance industry are artificial intelligence (AI) and machine learning. AI and machine learning help resolve claims exponentially faster, empowering teams to intervene at the right times, as needed. McKinsey goes so far as to predict that, by 2030, “claims for personal lines and small-business insurance [will be] largely automated, enabling carriers to achieve straight-through-processing rates of more than 90% and dramatically reducing claims processing times from days to hours or minutes.” Teams will be able to personalize solutions based on real data accessed instantly instead of requiring weeks of human analysis that can’t possibly incorporate the millions of data points that machines can. Technologies help claims representatives do their jobs faster and smarter so that people receive better care and organizations enjoy significant improvements in their combined ratios. Insurtech Becomes a Thing While AI-based solutions are relatively new to the commercial insurance industry, they have already given rise to a whole new market segment — insurtech. More and more companies are implementing AI into their offerings, and insurtech vendors are experiencing rapid growth of their customer bases as well as extending the contracts of existing customers based on the early benefits they’ve experienced. As the industry enters its biggest period of innovation in a century, driven by insurtech startups, and as organizations begin to track the magnitude of cost savings and other benefits, several important lessons have emerged. Below are my takeaways thus far from what’s becoming an AI revolution: See also: How Claims Process Must Drive Change   Don’t Do AI Just to Do AI There are cool applications for AI hitting the market all the time. Some sales guy is going to come in and show you something that will make your eyes light up. Don’t give in to temptation and sign on to do something just because it is an AI-based solution. Instead, think about your organization’s most pressing needs. What are your pain points? Where are your hold ups? What are employees frustrated with? If a solution fails to address these needs, it’s just another shiny tool that will never be used to its full potential. Practicality and usefulness are essential. Ease of Use Matters You could find the best solution in the world. It could be designed perfectly to take care of your problem, but, if it’s hard to use, it’s virtually worthless. Employees must want to use a new solution; they need to see how it streamlines their daily tasks and makes their jobs easier. This is essential for adoption. Adoption of AI-based tools can often face generational hurdles. There has been some resistance to AI out of fear that it will take over an employee’s job or simply from the desire to maintain the status quo due to personal comfort and familiarity. These are very real fears, and you would be wise not to discount them. The best way to prevail with a new solution is to show clear benefits to employees and make tools and software as easy to use as possible. Make a Plan AI requires forethought — not just in terms of what an organization needs or how employees might use it but also in terms of how it will be rolled out. There has to be a plan for implementation. Who is going to lead the project? How will employees be trained? What will happen with the data once it is generated? Even the best AI-based tools require management. Decisions need to be made in advance to get the most from any solution. Otherwise, implementation can lead to chaos and frustration, and the luster of a powerful new tool will wear off before it’s ever really put to use. Don’t Try to Do Too Much One of the biggest mistakes companies make once they understand what AI and machine learning are capable of is to take on too much. I would strongly encourage organizations to define and maintain a singular focus in applying the technology. When a company’s primary goal is to generate cost savings, for example, everything it does should turn in that direction. After all, the biggest advances come not when one goes broad but when one dives deep. AI applications are no exception. When organizations maintain a singular focus, they can devise the most consistently innovative and necessary solutions for their teams and customers. Real Personalization Is Possible Personalization has long seemed like a myth in the commercial insurance industry, something elusive that every company chases. With AI and machine learning, personalization will soon become a reality. See also: Why AI Will Transform Insurance   Because AI-based solutions can handle absolutely massive amounts of structured and unstructured data — and because they can learn on their own — users gain highly nuanced levels of information, which they can then apply to customize offerings. This opens the door for all kinds of opportunities to develop custom policies or benefits based on relevant data points. When personalization comes into play, everyone wins because costs, care and objectives are all aligned. Next Generations of Claims Operations On top of cost and efficiency benefits that AI is already demonstrating, the solutions of the future will improve the claims process across the board. For example, the need for litigation is reduced when claims are addressed in a timely manner or when injured workers get in to see the best doctor right from the start. Medicare Set-Asides (MSAs) can be processed in a fraction of the time based on better data. The possibilities are virtually endless when it comes to processes that can be improved. What we can see, however, even from these early days of insurtech, is that AI and machine learning will fundamentally improve how care is distributed and help the entire commercial insurance industry evolve. Looking forward to the years ahead. As first published in Claims Journal.

Read More