Personalized offerings for navigating rapidly evolving customer needs
Maintaining a high level of customer satisfaction becomes of utmost importance for global insurers as an overwhelming 80% of CEOs actively incorporate customer satisfaction metrics into their long-term strategies [2]. In their quest to attract and retain customers, developing the capabilities to analyze, predict and deliver what the customers need before they even request it soon emerges as a key competitive advantage. However, anticipating the customers’ needs has become increasingly challenging as they change rapidly and unexpectedly. Amidst such a circumstance, insurers are now turning to AI as a potential solution. Indeed, a survey by EY indicates that nearly half of the insurers plan to utilize AI for trend & demand prediction and operation optimization [3].
One prominent AI use case in navigating evolving customer needs is personalized insurance policies. Unlike the traditional approach, where customers are grouped based on demographic and risk-based criteria for calculating premiums and coverage, personalized insurance policies provide customized options that are tailored to each customer’s unique needs. AI’s capabilities to study a significantly large number of factors allow insurers to study their customers’ behaviors more accurately, hence adjusting premiums accordingly to improve risk management and better match customers’ needs. The application sparks enthusiasm not only within the insurers but also the customers, as 6 in 10 respondents show a willingness to share personal data in exchange for lower pricing [4]. Global insurers have been exploring the potential of AI in this regard, with the UK-based insurer Zego standing out as a prominent example. The insurer launched a so-called “intelligent cover” program that offers insurance services for gig workers such as Uber drivers, promising to lower premiums if they sign up for monitoring. Under this scheme, the company utilizes AI to analyze a number of inputs, including traditional elements such as age and more complex, real-time data such as fast braking and cornering, to identify its customers’ risk levels and adjust premiums accordingly [5].
Underwriting and Claim processing automation for cost optimization
Underwriting and claim processing are deemed to be time-consuming and labor-intensive, which hurts not only the insurers but also the customers. Reasons for such a case depend on the type of insurance, but one common issue is the complexity of data entry, which involves manually recording and verifying a myriad of customers’ details. As a result, it is estimated that administrative tasks such as manual data entry account for around 40% of a life underwriter’s working time [6]. Moreover, human errors seem to be a major contributing factor, as approximately 20% of data collected, processed and entered by insurance agents and policyholders are inaccurate [7].
Given these reasons, underwriting and claim processing are ripe for automation, with technologies such as RPA and AI being the key enablers. With automated underwriting, customers and agents no longer need to manually collect and enter data; instead, they can simply provide documents, forms or any data that can even be unstructured (such as hand-written documents). This information will then be extracted using AI-powered technologies such as Intelligent Document Processing (IDP) and then be imported into the insurers’ systems with Robotic Process Automation (RPA) bots. Once needed information is imported into the system, AI will analyze these data against pre-set criteria and thresholds, providing scores for each customer and comparing these scores with pre-defined rules to make decisions or inform human decision-makers.
This AI application has become increasingly popular in recent years, with global insurance companies incorporating it into their day-to-day operations. For instance, a multinational insurance and finance corporation headquartered in Central Hong Kong has partnered with FPT Software to implement underwriting and claim processing automation, leveraging the IT service provider’s Confidon solution. The Hong Kong-based insurer successfully automated their underwriting and claim processes, reducing the processing time by a whopping 600%, from 20 minutes per policy/ claim to merely 2 seconds. Moreover, the solution also aided the insurer in improving sales and risk management by identifying customers with high fault and lapse risks.
Learn more about the success story here.
From “Detect & Repair” to “Predict & Prevent”Climate change is becoming one of the most concerning threats to the insurance industry. Research found that in recent years alone, the world has lost over US$ 1 trillion in damage and suffered 1,000 extreme weather events. Consequently, the insurance industry experienced a spike in claims for natural disasters by 54% in 2022, compared to the most recent 10-year average [8]. To make matters worse, climate change can be the leading cause of other issues, such as illness and diseases. In these circumstances, the global insurance industry is shifting away from the traditional “detect and repair” approach and towards the “predict and prevent” model, meaning that insurers actively work with their customers to prevent claims altogether. In their quest to accelerate this transition, AI has soon become a key enabler.
Several global insurers have begun utilizing AI to predict, inform and subsequently help their customers prevent incidents. For example, the Swiss insurance company – Zurich leveraged an AI neural network to create a hazard map for the current and future climate, combining data from climate models and the current hazard map [9]. Other insurers are taking a step further by incentivizing customers to take precautionary measures. For example, the UK-based insurtech company – By Miles has been actively developing processes to warn customers of high weather risks and encourage them to move their vehicles to the high ground [10]. Similarly, Ping An, one of China’s biggest insurance groups, offers an AI-powered diet and exercise plan to its Type-2 diabetes policyholders to reduce their chance of developing complications. The technology is trained on data on diabetes complications and then analyzes the policyholders' health through an app or a blood-glucose monitor. Lowered premiums at renewal are promised to customers who commit to following the company’s diet and exercise plan [11].
Beware of AI bias
Despite the outstanding results provided by AI, insurers should be aware of the so-called “AI bias” phenomenon when AI algorithms provide biased results. Sources of AI bias can vary, but some common causes include insufficiently diverse training data, biased algorithms, or cognitive bias caused by human developers’ experiences and judgement. Successful AI implementation is only possible when insurers are aware of and map out a careful AI governance approach, and working with a highly experienced AI partner is one optimal solution for accelerating their AI journey.
