The insurance sector is undergoing considerable disruption – a process described as “continuous revolution” and “the new normal.” There is plenty of scope for both start-ups and incumbents to innovate within that, both in domestic and emerging markets.
In this piece we are going to look at some of the key insurance industry technology trends, and where they are taking us.
Using data to move from passive risk amortization to active insurance risk management
The internet-of-things (IoT) revolution means that a huge volume of data is now available from sensors in vehicles, home automation systems, and mobile devices. The use of IoT devices allows insurers to develop new products. For example, insurers can provide individuals products like travel health insurance good for a fixed number of days abroad. Using the customer’s smartphone with GPS, the insurer can start the clock on the policy when the customer leaves the country, stop the clock when the customer returns to their home country. In industries that use shipping, we see a desire to charge for insurance on shipping for goods and the vehicles that transport those goods based on where the vehicles go. If the vehicle stays in clear weather, away from war zones, and on safe roads/shipping lanes, then the insurance should be priced accordingly. IoT allows insurers to monitor behavior and build software to price in the new risk. The insurer can also inform the user what their actions mean so that the individual or company can act to mitigate changes in risk. IoT, combined with machine learning models makes this possible.
The same data that allows the insurer to move towards an active insurance risk management model also powers a more personalized insurance model for the consumer. Currently, we use gross premium classes based on age, gender, and location. Data science techniques have the potential to allow individually tailored, lower-cost premiums, while also managing the overall risk in the book.
Underwriting and distribution
As more of the underwriting decision-making process can be automated using personalized insurance risk-modeling, the underwriting process itself can be automated. Underwriting automation creates an opportunity for the elimination of the traditional agent and direct-to-consumer distribution.
Speeding up the process can have a considerable impact on a consumer’s likelihood to purchase. Deloitte, for example, has found that, in life insurance, a customer’s propensity to buy increases from 70% to nearly 90% as the decision process tends towards to a real-time, interactive experience.
By providing direct-to-consumer channels, agent commissions and other costs can be eliminated. When up to 80% of the premium is currently taken by distribution costs, this can help protect margins, lower premiums and enter new markets. Bima, for example, is offering low-cost micro-insurance products in emerging markets.
Another area where there is a considerable amount of human effort in the insurance value chain is in claims processing. Here, the opportunity lies in putting that process in the hands of the consumer. Advances in cognitive technology like voice and image recognition and language understanding allow us to create so-called “bots” which take the customer through the claims process. This can be used to automate much of the decision-making process. The first generation of this technology is being developed by companies like Rightindem, but there is still a great deal of scope for innovation.
Big Compute, Big Data, Big Math
A dependence on sophisticated modeling, machine learning, and cognitive services drive all of these insurance industry trends. These are technologies that consume a considerable amount of computing power.
The public cloud has democratized access to high-performance computing, both directly by developing applications to run in the cloud, and indirectly by (acquiring modeling-as-a-service products from actuarial technology providers like Milliman, FIS, and WTW).
As personalized insurance modeling grows, so will the demand for compute. That creates a challenge for both infrastructure providers like Microsoft Azure and application developers to optimize their use of resources. One way to deal with this is to explore technologies such as compute-on-GPU and compute-on-FPGA.
Compute on GPU uses the fact that GPUs are designed to optimize mathematical operations on streams of data – originally to model and render 3D scenes on 2D display devices at a high frame rate – to deliver general purpose computation at high speed and low energy cost.
Compute on FPGA allows modelers to take a core part of the algorithm that needs to be optimized and implement it directly on a chip – a kind of “programmable silicon.”
The rate of innovation in the sector offers both risks and benefits to incumbents. The risk is that disintermediation and direct-to-consumer offerings from disruptive new players act as a “death of a thousand cuts” and leaves them holding only higher-risk products on their books. The opportunity is in leveraging what is known as the “API Economy” to integrate new technology from third parties into their value chain.
We also think that APIs provided from outside the sector, such as the Open Banking initiatives, and home-automation IoT standards will become important in the modeling and decision-making process.
Seed capital in the insurance sector is coming from a mix of traditional VCs and private equity companies, and the incumbent insurance players themselves. We expect this trend to grow over the next few years, and for M&A activity to increase as a result.
This allows the big players to leverage their expertise and brand capital and rapidly experiment with bringing new products to market, without significant investment in the underlying technology, and turning the competition into partner organizations.
This article presented current insurance technology trends. If you would like to know more about how Azure can be used by the Insurance industry, visit Azure for Insurance page.
Nick Leimer is the Principal Insurance Industry Lead on the Azure Industry Experiences team led by Paul Maher. Nick brings 20+ years of experience bridging the business and IT gap in application development, operations, and infrastructure. Before joining Microsoft, Mr. Leimer was the Senior Director for Actuarial Compute at Manulife leading the migration of LAN based HPC Farms to the Azure environment for all of Asia; including all desktop Moody’s / GGY Axis users and data feeds. Prior to Manulife, his career progressed through developing applications for Actuarial Valuation and Projection including components of ArcVal and Prophet, the Defined Benefit area with the Pension Benefit Guarantee Corporation (BPGC), and leading life insurers, AIG and Manulife.