How to bring efficiency to the insurance claims process

2020 marked the beginning of a tumultuous chapter for the insurance industry. A global pandemic changed the way we worked and forced carriers to rapidly develop new remote work strategies. "The Great Resignation" resulted in tremendous amounts of knowledge leaving the industry. And, a general lack of interest in insurance by younger workers left insurers with a wide gap to fill. 

It is no surprise then that a 2021 Jacobson Group/Aon study found that staff expansion throughout 2022 was top of mind for more than half of survey respondents. While many of these understaffing challenges are a result of long-term, experienced employees taking new opportunities elsewhere or approaching retirement, knowledge gaps are left unfilled across organizations that are unable to keep the pace of hiring and training up with customer demand. In this current environment, it may simply be seen as impossible to match the pace of hiring and training to that of those leaving the industry. 

The good news in all of this is that perceptions are not the same as reality. Despite an, at times deserved, reputation for taking a conservative approach to technology adoption, insurers view digital transformation strategies as one important way to bridge the knowledge gap. But despite best intentions, technology adoption is not a panacea. Digital transformation must be approached holistically, or these crucial efforts are destined to deliver suboptimal results. For example, deploying incompatible systems that are not interoperable with one another creates redundancies and workflow gaps, overworked and overloaded employees and operational inefficiencies. 

To avoid these pitfalls, we must evaluate how to best embrace digital transformation. The adoption of a modern technology stack, that incorporates intelligent decisioning, will enable insurers to deliver the speed and accuracy of decisions incumbent upon carriers, operate with greater efficiency, and ultimately deliver exceptional customer experiences.  

The claims process: A microcosm 

One need only look at the claims process to understand how an intelligent decisioning-based approach to digital transformation can benefit both the insurer and the insured. The traditional claims process has long been highly manual and driven by the claims professional's knowledge, experience, and expertise. It is also driven by hundreds, if not thousands, of micro and macro decisions made individually or collectively. And although there has been technology successfully applied to claims (think claims management systems such as Duck Creek or Guidewire), most digital transformation attempts result in tenuously connected disparate point solutions which fundamentally address individual aspects of the claims lifecycle (e.g., claims management) versus a more holistic approach to the entire claims process from FNOL to final settlement. Furthermore, these attempts to apply technology to the claims process are typically not focused on helping claims professionals make better decisions, but rather ensure that individual claims are appropriately tracked through the process.

But what do we really mean by this? Let's look at a common claims scenario that could benefit from an insurance decisioning approach. A claim is filed for a two-car accident. Although the vehicle did not incur significant damage there are reported injuries. A trained and experienced insurance professional reviews the claim and determines that the reported injuries could be consistent with the reported damage and that in the absence of other mitigating factors recommends that the claim should be settled. Flagging the claim for further review and investigation when it is not abundantly clear that there is anything wrong can add significant time between FNOL and settlement, add workload to an already overworked SIU and potentially damage the policyholder relationship. It simply feels that the most prudent decision is to settle and move on.  

Admittedly, the scenario described does not employ a significant amount of technology. But let's look at it again with the assumption there is at minimum a claims management system involved as well as rules-based fraud detection in place to help review claims. In this case, the claims handler may receive an alert that a risk factor was triggered - minor vehicle damage with reported injury - but again decline to forward for further investigation. It is decided that a single, inconclusive flag, that most likely would have been identified even if technology was not involved, is not enough to trigger an investigation.  

Adding decisioning to the stack 

Now, let's examine our same scenario through the lens of insurance decisioning. Technology adoption, digital transformation initiatives and automation are already rapidly changing how insurers think about the claims process. An increased focus on policyholder satisfaction and a need for greater operational efficiency, as previously discussed, are key factors driving insurers' strategies in this area. But as we have learned from previous attempts at automation, automation for automation's sake will rarely be successful. Automation must be purposeful and implemented with an understanding of what can, what cannot, and what should be automated.

So how does this change our two-car accident from before? Let's make some new assumptions, primarily that the insurer has made investments in claims automation and fraud detection as well as the use of artificial intelligence. At FNOL the claim is reviewed by AI to determine if it may be a candidate for straight-through or expedited processing. The policyholder has submitted all the proper documents, corresponding photos are legitimate, repair estimates are in line with industry norms and even the medical costs associated with the reported injuries are acceptable. However, the claim is not recommended for expedited settlement and instead is referred to a claims professional for further review. 

Unfortunately for the policyholder not only is the initial red flag present - minor damage with injury - but also the AI has picked up some further inconsistencies that may not have been obvious even to an experienced claims professional. AI-based entity resolution shows that the policyholder, using different permutations of names, addresses and other PII has made questionable claims in the past. It has also uncovered multiple network connections between the two parties involved in the incident and service providers associated with the claim (lawyers, repair shops, medical, etc.). And, although the accident happened 100 miles from the policyholder's home address, both participants live around the corner from each other. Now the claims handler has everything at their fingertips to make the best decision possible about what to do with the claim and arm the SIU with all the relevant information they need to begin a proper investigation.     

As we have explored, even the most common types of claims may contain complexities that are not readily apparent to human claims handlers. Further, simply automating aspects of the claims process does nothing address this and can, in fact, exacerbate the situation as more and more experienced professionals leave the industry.  That is why the use of artificial intelligence and machine learning to bring intelligent decisioning to the insurance industry must be the next paradigm shift. And it is not only the claims process that can be positively impacted. Arming claims professionals with the right tools to make the best claims and underwriting decisions possible helps insurers bridge the knowledge gap being created by current employment trends, operate more efficiently and cost effectively, and perhaps most importantly deliver exceptional policyholder experiences.