How is technology impacting the insurance industry and what do we expect to see in the years to come? To find out, we sat down Alice Underwood, Global Leader of Insurance Consulting and Technology and Dave Ovenden, Global Pricing and Underwriting Leader for their take.
How will automation, AI and big data impact the insurance industry?
Dave: I think there’s a range of outcomes depending on the nature of the portfolio and how you define those terms. A good place to start is the overuse or broad use of the term AI, which isn’t just about well-trained computers, but encompasses a huge breadth of technology from deep predictive models to cognitive learning. A lot of investment has been in predictive modeling and that will increase as cognitive computing capabilities expand.
As for big data, our view has long since been that it starts at home, so having a good grip on your underlying data assets so you can bring in the capabilities from AI predictive models and deploy them against those assets will be transformative. And I think there’s a really good opportunity to think about how this technology could play out differently in personal lines – where products are simple and the pricing is sophisticated – versus commercial lines, where products and exposure are very complicated, but the rating is less sophisticated, and there’s less data to work with.
Alice: As Dave says, we need to step back and think about what those terms mean. Automation is about getting a computer to do a well-defined set of tasks, whereas AI can take on almost anything that we can get a computer to do and follows an iterative learning trajectory. But it’s not an either-or choice: automation and AI can work together. Automated systems can call upon an AI solution for insights, either to be used in an “intelligent automation” process or to be handed off to a person.
How is technology helping insurers increase growth and profitability?
Dave: Insurers are using data from predictive models and other types of AI to build mobile apps that drive the right product (coverage, limits and deductibles/excesses) to the right customer. Predictive models are no longer just about pricing and underwriting rules, but are being deployed in sales and operational effectiveness too. We’re also seeing the emergence of electronic trading hubs and personalized aggregators that allow customers to compare policies and rates, a trend that will continue to expand into less complex commercial risks. I think some really interesting interactions will start to occur when brokers and intermediaries start using technology to make decisions about which insurers they use and which deals they send to their customers.
In the near future, we’ll see commercialized trading hubs that have intelligent automation driving sophisticated pricing into the hub from the insurer’s side – and from the broker’s side, automated decision rules about what to present to the insured. So there’s really interesting things around the automation connected to distribution. Insurers that connect to these hubs, and continue to successfully manage distribution relationships, will see growth because it will allow them to quote for almost zero marginal cost, while mitigating the risk of being commoditized by aggregators.
Right now in the commercial market there’s a lot of human intervention and not a lot of data integration; but once that changes, we’ll see a reduction in cost through the ability to make better decisions quickly. Within the commercial community today there’s a lot of frictional cost in the value chain. As a result, there’s significant opportunity to employ intelligent automation to augment the underwriter’s decision, so they can choose which risks run through a no-touch process, which go through a low-touch process and which need a lot of underwriting intervention. This will lead to much more value around what an underwriter actually spends their time doing.
On the personal lines side, companies that are on top of data enrichment and have a culture of finding new novel sources of data c – along with an understanding of their own data and how to deploy it – will be successful.
Alice: There’s tension between growth and profitability. It’s easy to grow quickly if you’re willing to lose money… and it’s easy to set well defined parameters around how profitable you want to be, but it may be difficult to grow under those circumstances. Technology can help you balance those competing priorities and explore options to find the right place on the risk/reward curve. It can also help you expand your range of products and provide them to customers quickly, whether that’s through apps on a mobile device or through an online portal, both of which can help you grow and access your customers more efficiently and cost effectively.
How will technology address the range of regulatory environments in insurance?
Dave: Technology has to be sensitive to regulations. We work in environments that move from a complete tariff, to those where rules and prices are disclosed to regulators, to fairly deregulated markets such as those in the U.K., where it’s more principles-driven than framework driven. So, technology needs to support all of those markets.
In less regulated markets, technology can help you understand your segments in a much more granular way so you can be much more focused on what you charge. In the U.K., an example might be the young driver market where telematics are very heavily used, so the way for a young driver to enter the driving population is often through a telematics product. That’s a really good example of regulation allowing the market to respond to a population that represents something like 5% of the community and 30% of the losses.
Alice: Given the range of regulatory frameworks, there are a number of different strategies insurers can deploy. Where the rate environment is flexible, insurers may want to focus technological development around pricing, whereas those in regulatory environments where pricing is relatively constrained may want to focus technological development on underwriting or claims. In terms of managing claims, you could automate claims processes to be more efficient and cost effective. You could also use artificial intelligence in the claims process to bring the claims that need attention to the right handler and process the claims that don’t need close attention in a no-touch way.
How is technology affecting the average insurance buyer?
Dave: I think it will be different for each segment of the market. In personal lines, we’re already seeing greater transparency in products and pricing, particularly in the U.K. I think we’ll start to see products emerge that link people’s ability to work with business interruption, so there’s a more holistic product for small businesses. As a result, personal and commercial communities will see more tailored products.
Alice: Products for individuals and small businesses are becoming more customized and there’s a much wider range of options, but that can make choosing the right coverage bewildering – which is why many customers are using automated advisors to aid their decision making.
How do you see insurance changing over the next decade?
Dave: The technology around handling unstructured data is an interesting area for commercial insurance. There will be greater ability to consume printed material and less structured forms of data and incorporate that information into intelligent automation in a way that we previously haven’t been able to. Ingesting reports (e.g., claims, medical and survey and engineering reports) and turning them into actionable data and instantly usable insight will come together in the next couple of years. That will lead to an ever-richer source of data for models because you’re no longer bringing just facts and figures together, you’re bringing insight to unstructured data.
The issue of privacy will run in parallel as people and regulators begin to think more carefully about it. I could see an environment where personal data isn’t given freely, which insurers might have to cope with and adapt to. We’re already thinking about who owns the data, and that’s especially relevant as we’re sending and receiving more data between parties trying to get insurance for a customer. We could also see shared data assets, particularly around commercial exposure data, one agreed asset register rather than re-keying the same data many times across the insurance value chain.
We have data aggregators, data collectors and people making money along the data value chain – except for those at the beginning, the original data owners. I’m not sure that mindset is going to persist into the next decade.
Alice: I think we’ll continue to see increased connectivity in the market. Companies will form alliances and partnerships with those previously regarded as competitors, and with players they just hadn’t considered before. It’s no longer about vertical integration. It’s more about being able connect to different providers of information and technology and integrating that technology with legacy systems.
Alice Underwood is Global Leader of Insurance Consulting and Technology at Willis Towers Watson
Dave Ovenden is Global Pricing and Underwriting Leader at Willis Towers Watson