Skip to content

Digital disruption in the insurance industry: Leveraging big data with flexible “greybox” algorithms

Update December 2017: TMNS is now Devoteam

The insurance industry has frequently been said to be conservative and slow to take in technological developments: Insurers often do not leverage the full potential of big data and advanced analytics to make business practices more efficient and effective. Meanwhile, the growing threat of nimble and agile new entrants in the insurance space is placing ever increasing pressure on existing business practices, as they do know how to leverage data and technology. Whereas incumbents struggle to move entrenched bureaucracies, change established work practices, and add functionality to outmoded legacy IT systems, start-ups such as Insureon and Lemonade are disrupting the insurance marketspace by introducing new and more cost-effective insurance products and services. In short, it appears that the insurance industry is ripe for a broad-based digital disruption, and that incumbent insurers can no longer build incrementally on their current modus operandi. To be successful as an insurer in attracting and retaining valuable customers, innovation is now a matter of urgency.

Digital transformations on capital markets

To see what a digital transformation that fully builds on data and algorithms looks like, one need only look at how capital markets incorporated electronic trading. While new information technologies had slowly been making existing business processes run faster since the early 1980s at NYSE, its merger with a high-tech startup in 2005 led to a cultural sea change after 214 years. Following about a decade of transition, product pricing engines are now measured in milliseconds, relevant information fully impacts market prices in seconds and days rather than weeks and months, the market has become more transparent, and transactions have become more cost-effective for consumers.

Building on this transparency and speed, market players have also created new product categories (ETFs) that allow end-users to invest directly in countries, industries, and commodities, at incredibly low costs and high speed, even on very large transactions. Changes at the exchange have also had their profound effect on participant organizations. UBS and Goldman Sachs are frequently cited as examples of algorithm-based efficient work practices: While UBS once owned the largest trading floor in the world, this was nearly deserted some time ago and is now for sale. Goldman Sachs now runs one of its departments with 2 persons rather than the 600 that it once employed, running the same or more activities and being significantly faster to respond to market events. Such data and technology-driven business practices in financial services have been dubbed FinTech; Transferable knowledge, concepts and methodologies that Devoteam has ample experience with.

InsurTech at Devoteam

InsurTech is an insurance-specific aspect of FinTech, and aims to introduce data and technology driven approaches to the insurance industry. New and more cost-effective insurance products and services incorporate our newfound ability to gather and process big data into the actionable intelligence that drives business processes: Competitive advantage in the insurance industry will come from the most effective use of data and technology. Broadly, this can mean that predictive analytics can be incorporated into business practices, to fuel artificial intelligence-based customer interactions. Such disintermediation implies that advisory intelligence can be provided as robo advice for an improved cost-effectiveness, broader reach, 24/7 availability, enhanced scalability, and more consistent offering based on corporate principles. Another example of disintermediation are new product categories, such as peer-to-peer insurance models.

Furthermore, by utilizing the latest advances in big data and advanced analytics, InsurTech provides the ability to go beyond generalized linear modeling of risk: Rather than relying on broad actuarial tables, more precise micro-categories can be established to determine appropriate risk-reward and pooling. In addition, the actuarial function can be enhanced, augmented, or in some cases even replaced by artificial intelligence. A sign of our times is the increasing production, quality, and ubiquity of data such as that espoused by the Internet-of-Things and wearable biometric devices – we are creating a data ecosystem unlike anything the world has seen so far. Building on big, fast, and sometimes even streaming data, more accurate risk insights can be leveraged to provide for dynamic pricing models. One can even imagine micro-event insurance offerings, such as when one borrows a friend’s car.

This data driven approach to business processes builds on predictive analytics that allow the organisation to understand what the likely impact of events, offerings, and courses of action may be. Examples of this include legal claim estimates built by algorithms, health insurance quotes based on data coming in from wearables and a data based medical history, accurate legal claim value estimations, and car insurance quotes based on driving habits. Beyond quotation systems, predictive analytics can be leveraged to foresee equipment failure and act before the damage is done, or sense and take preventive measures on impending health problems. Building on streaming real-time weather data, crop insurance pricing can now be based on dynamic patterns.

Finally, customer offerings can be enhanced by an improved interaction model with customers, such as through mobile apps. This improves accessibility to products and policies, potentially also allowing for customized and pro-active offerings from recommendation engines. Such relationships can be enhanced by more frequent (artificial intelligence-based) interactions, effectively building a deeper trust relationship between insurer and customer.

The greybox proposition and the insurance industry

The algorithms and architectures that power the preceding scope of examples are most aptly leveraged by building and tuning the appropriate interactions between human resources and technology. Firms often rely on blackbox functionalities as these purportedly offer the most consistent, scalable, and reliable business process. Unfortunately, this often leads to a front-end “computer says no” scenario, where personnel is unable to adapt offerings in appropriate circumstances. In extreme cases, such as on capital markets, algorithms-gone-wild have even led to near instant bankruptcies, for instance when Knight Capital lost $440 million dollars in 30 minutes.

Insurance companies can become agile and data driven organizations by implementing artificial intelligence and advanced analytics in business processes the right way, thereby avoiding algorithmic rigidity or even disasters. At Devoteam, we leverage years of experience building greybox systems, designed to leverage the contextual awareness and flexibility of human capital with the speed and discipline of algorithms. We are not only able to guide the insurance industry in leveraging big data and artificial intelligence, but also have ample experience in fine-tuning the business processes to create agile organizations enhanced by greybox systems.

Rather than building on slow and expensive processes based on human analysts processing information, insurance companies should build algorithmic artificial intelligence-based business practices. By doing so, insurers can deliver new and more fine-tuned products faster, with a more timely and accurate understanding of risk, with better customer accessibility and deeper relationships, whilst potentially driving down overhead costs by an estimated 75% to 95%.