(Image credit: Shutterstock.)
InsurTech remains a growth sector in 2019 with new entrants in the market. Funding continues to flow to start-ups albeit at a slower pace than 2017 and 2018. CIOs and CEOs face an enormous array of product choices, each promising to deliver competitive advantage. Overlooking a game-changing technology may adversely impact profitability or provide one’s competition a chance to gain market share. Insurance executives are also concerned with the amount of change and disruption that accompanies the implementation of new tech and how their businesses will handle it. I recall the first half of the decade when insurance executives faced with up to a couple of dozen choices in core systems, turned to consulting firms to assist with making the right selection for their firm. Some experienced failed implementations and others were able to implement successfully only after a prolonged and protracted effort. No carrier has implemented a core system without facing some disruption.
The situation today is more challenging with carriers facing an onslaught of new technology such as artificial intelligence, NLP, chatbots, IoT, blockchain, big data, etc., to name a few. There are literally hundreds of vendors to choose from and there is no consulting firm that is familiar with all these choices or has the necessary expertise to guide a carrier client through this line-up of InsurTech “must-have” technology. Matters are made more challenging with the market entry of InsurTechs that sell insurance instead of technology. There is also the matter of core systems implementation fatigue to be considered. So what’s a CEO to do?
Industry publications, analysts and other experts have written about chatbots that improve agency and policyholder experience; predictive algorithms in underwriting/rating; artificial intelligence in claims automation and severity management; drones for property loss data collection; image processing to generate auto damage estimates and artificial intelligence in loss control/risk management. Insurance companies are transforming digitally, and changing how risks are rated, how insurance is sold and how services are delivered to policyholders.
Analyzing the Insurance Value Chain
At Ajira AI we examined the insurance value chain to identify the business functions that weren’t fully addressed by core systems or where there were clear gaps in core systems functionality. A deeper look at the edges of core systems functionality helped us identify unaddressed areas such as claim lag. In addition, we attempted to determine areas of strategic change based on prior industry experience and insight. These are areas we felt would be most likely impacted by a combination of demographic, technology, service level and regulatory shifts. We challenged our formal analysis and findings by performing breakdowns by function and by external/internal touchpoints. We were able to identify several areas that Ajira AI is currently targeting with new products that blend artificial intelligence, speech recognition, mobile and cloud technologies.
One of the areas that workers comp and other commercial lines core systems are unable to address adequately is claim lag. Collecting data at the point of injury or loss, during or right after the incident occurs remains a challenge. Industry wide this problem, referred to as claim lag, may cost a carrier up to 4 percent of its written premium annually. Our IntelliFROI and IntelliFNOL products address this industry problem by providing a smartphone capability that captures injury/loss details at the point of occurrence including demographic data, photo, video and injured party/witness statements. Our products use a combination of AI and speech recognition algorithms to pre-populate policy data and increase accuracy of the reported injury/loss information. We also check for fraud and subrogation opportunities as the incident is sent to either a policyholder portal or claims system for further processing. Our products use state-of-the-art engineering to maintain up to date regulatory compliance.
While InsurTechs are focused on commercial and workers comp front-end distribution, agent/policyholder experience, predictive risk scoring/rating, intelligence in claims automation, etc., our analysis has led us to identify fraud, subrogation, loss control and premium audit as areas that are ripe for application of AI algorithms.
Finally, our analysis has made clear that it may not be sufficient for carriers to rely on their corporate databases to provide the attributes and the quantity of data needed for machine learning and artificial intelligence. Carriers that are able to anticipate change and incorporate it into their data collection and transaction processing early will benefit from better outcomes of machine learning and artificial intelligence algorithms. While some data may be available for purchase from third parties, there is definitely a lot more that a carrier may achieve by thinking outside the box or as author Tom Friedman puts it—thinking without a box.
Recovering Lost Value through AI Algorithms
For example, in workers compensation it’s been known for many years that non-English speaking workers have accidents at a higher rate than English speaking workers. However, this data is not collected or used during underwriting nor is it captured during the claims process. This is a potential opportunity. The legal and regulatory framework in the various states will evolve in time to allow for this data. Regardless of when English/non-English speaking bi-furcated rates for payroll or associated changes in rating algorithms are adopted, there is no reason why carriers should not be capturing this information during the claims process today. This data has several uses, for example it may be used in risk scoring during renewal ahead of broad adoption by the individual states or by competition, assuming of course that it’s legal in the state where the policy is being underwritten. The data may be used by loss control and risk management to tailor and delivery safety training. The data collected by these functional areas may be made available to premium audit to perform a check on payroll classification. These attributes may also be used in fraud identification. Therefore, a single attribute may be used in multiple business areas across the insurance value chain. This is just one example of the type of thinking needed to recover lost value through the application of AI algorithms.
It is also known in workers compensation that new employees have accidents at three times the normal rate in their first month. Are carriers able to influence policyholders to provide new employee start dates for higher risk class codes so that better safety training may be provided? Health insurance companies underwrite individual health today. It’s not inconceivable that at some point carriers writing workers comp will be able to ask for and receive lists of employees and contractors as part of application underwriting. This will enable carriers to offer a custom product to each policyholder that is uniquely rated based on their specific workforce and risk. It will allow carriers to identify employees/contractors that may have multiple claims in their past and rate for such risks appropriately. It will dramatically lower costs for those employers with a good safety record and a safe workforce instead of the current rating method that pools all loss experience for a job class into a rate for an entire state. Core systems and current predictive rating models that are proprietary and closed and will need to be open and flexible to accommodate significant shifts that the workers comp and other commercial lines are going to undergo in the future. After all data is gold. The good news is that when one examines the insurance value chain in commercial lines including workers compensation, there are many such opportunities.
To learn more about Ajira AI, click here.