Due to the progress made in the development of artificial intelligence (AI) and automation, the insurance industry can now take a closer look at an area that was previously not in the focus of digitalization: underwriting.
Whether health, life, property or commercial insurance: the damage risk is assessed before an application is finalized. A decision by the underwriter should consider the applicant’s risk profile appropriately; this highly-qualified advisor tries to maintain the highest possible degree of objectivity. The underwriter does a balancing act between the two poles of experience and objectivity – though they may not realize this. If somebody has frequently experienced that certain circumstances lead to a benefit obligation, this knowledge will be applied to a case that appears similar so that there is a risk of making a biased decision. If the risk appears to be too big for the insurer, the applicant will receive a rejection, which represents a decision that the affected party may frequently not understand. In the worst case, this means the end of an otherwise existing customer relationship for the broker. The digitalization of the underwriting processes solves this problem and offers even more advantages.
How digitalization benefits underwriting
The task appears clear and understandable: the purpose of underwriting is to efficiently insure the largest possible number of customers at an acceptable risk. In practice, however, this is an extremely complex process where efficiency is the key term. The representative and broker are on the front lines with the customer. After comprehensive consultation, a decision is made for a health insurance. The necessary health questions are answered and, ideally, the application is sent to the insurer in digital form. Unfortunately, the underwrite sometimes has inquiries. This is when a communication relay normally begins.
Tele-underwriting, as it is successfully done by different companies, relieves the broker and increases efficiency. Because through the software-guided process, no questions are forgotten, the decision on the application is made directly and all required data is directly available in the insurer’s system. Underwriters are given more flexibility to dedicate their time to more complex cases. The decisions from the software are based on deductively developed rules with such automated processes. Machine learning and artificial intelligence systems add to the software possibilities through the predictive options. „Predictive Analytics“ can adjust the rules for underwriting quicker on the basis of different data and information. The general predictions can be applied to individual cases as well, which represents a key along the way to individual rates in life and health insurances. Service providers get their AI system data from different sources – for example, from fitness trackers, weather analyses or also from social media posts – to create risk assessments and predictions. This is significantly more information than a person can process.
Ensuring access to voluntarily provided customer information
In order for AI to provide support when creating an individual risk profile for health and life insurance, the systems need data about the applicant. The more data the better. Information that many insured individuals already provide voluntarily: fitness wristbands, smart watches and smartphones record data along the clock, which in turn is relevant to evaluate their health. This information is not sent to the insurance companies, but rather to technology companies abroad. With “Health” and “Fit”, Apple and Google have created central apps in their respective user systems that aggregate such information. Consequently, Google recently strengthened the health area in their plethora of data by taking over the company Fitbit, which offers fitness trackers. The result is that the technology companies are currently better capable at evaluating the characteristics and therefore risk profile of a person than the insurance company. This is also the case to a smaller extent for vehicles, because smartphone tracking also offers a summary of time spent in a car.
The data sovereignty and data excellence of technology companies ultimately threatens the insurance company’s’ customer access. Apple regularly emphasizes that they do not want to use this data to directly earn money. But the initiative from Google to offer an individual bank account shows, using the banking world as an example, how quickly the ambitions of an IT company can threaten a business model. It would therefore be fatal for the insurance industry to have to be dependent on these companies to buy the necessary information for risk assessment and calculating rates. Or if they would be forced to the role of a supplier for IT companies that then offer their customers their “own” insurances.
Data, data, data
In order to minimize risks regarding these threatening dangers, insurance companies should work in two directions. On the one hand, measures must be taken to ensure customer access. There is still time to create individual offers, to establish corresponding know-how and to gain information about customers through cooperations with InsurTechs and other companies. The company Fitsense from the USA developed an analysis platform that is offered as a white label solution. The objective is to give insurance companies the possibility to establish their own fitness tracker or incentives. On the other hand, further efforts for investments in AI systems appear to make sense in order to be able to fully take advantage of the benefits that arise from predictive analytics. In this connection, AI forms the basis for objective risk assessments and individual risk profiles or rates. Underwriting supported by AI does not just go faster, but rather it is also ultimately less error-prone due to the large data quantities and machine learning.