How real-time data could close the risk knowledge gap

Talented photographers can tell remarkable and emotional short stories with a snapshot. A picture can be worth a thousand words, but a single frame is open to interpretation and finite in its ability to weave more complex tales. Movies provide a more robust storytelling canvas. But these image reels are still confined to a set point in time, and what was once viewed as innovative and amazing can quickly become outdated.

Data tells a story too. And how, when, and how often you can access the data determines how much of the story will be revealed. A single snapshot can only tell you so much, and long interludes between updates can leave vital information uncovered.

Today, more and more organizations are informing business decisions and guiding customer interactions with narratives gleaned from big data. The insurance industry is based almost entirely on data collection, interpretation, and application of big data stories to assess risk. But those stories are often fragmented, inconclusive, or provided by an unreliable narrator.

Access to real-time data allows insurance carriers to own the story and close the risk knowledge gap. Without the need for forms or additional customer outreach, missing and augmentative information becomes available with a simple search. And machine learning models can provide additional predictive intelligence for a complete business record in seconds.

Real-time algorithms replace static snapshots with a live feed for risk assessment. As stated by computer scientist Edsger Dijkstra, “A picture may be worth a thousand words—a formula is worth a thousand pictures.”

Big data, bigger picture
Knowing and understanding your customer is still the most essential component of sustained business success. The market cadence for commercial insurance is generally annual, with limited customer/agent interactions. A lot can happen to a business within this time, including adopting new service offerings, increasing the number of employees, or taking on new risks.

Real-time data supports real-time underwriting. A continual flow of data, instead of sporadic snapshots, allows for regular personalized adjustments in pricing and product offerings. According to a recent McKinsey & Company survey of small and medium-sized enterprises in Germany, 21% of small commercial-insurance customers switched insurers in 2020 due to changing risk landscapes. Taking a real-time approach to data collection allows carriers to keep better tabs on current and potential business, provide an individualized approach to service positioning, and reduce churn. The result is more focused and intentional outreach that results in higher written premium, lower acquisition costs, less turnover, and a better consumer experience.

Additionally, real-time data feedback can inspire insureds to voluntarily reduce risk. For example, telematics data for auto insurance has gamified the process of providing premium discounts by rewarding safer driving. Safer driving means fewer claims. Other carriers challenge insureds to walk a certain number of steps per day for discounts on health or life insurance, as daily exercise has shown to greatly reduce certain health risks. Gamification is much more effective than instruction and education in changing human behavior since it provides a more tangible and immediate benefit. And gamification within insurance is only possible with real-time data.

The constant underwriter
So, from where is real-time data sourced? How is it collected and categorized? It depends. As an example, 99% of the data that Planck collects comes from the open web, which are sources that are available to the public. There are also unstructured and meta data sources, as well as government repositories, which are not available through traditional search methods.

As any underwriter will attest, the internet is not structured to support insurance research. Pulling together a complete and accurate risk assessment is a complex and difficult task. Big data provides an opportunity to instantly know more about a business than ever before. And since this massive resource is constantly growing and evolving, processes for collecting and consuming this data have evolved as well with the emergence and advancement of AI.

Beyond a better consumer experience, the most apparent benefit of this method is the removal of human errors and omissions. Real-time data collection effectively closes the existing gap between claims data and underwriting data. Claims data is usually more reliably accurate than underwriting data, because adjusters and other entities are present to validate. On the other hand, underwriters are often forced to base their assessments on stale or insufficient risk data, resulting in potential premium leak. Rather than a snapshot, real-time data provides a continuous flow of up-to-date risk data for underwriter risk research with a higher degree of accuracy.

The truth is out there
A business doesn’t need to have a website to amass a digital presence. The online and connected nature of modern commerce creates detectable and collectable levels of digital exhaust for almost any organization. Consumer reviews, social media posts, videos, photos, and government documents can be analyzed to verify services, locations, and operating hours, and build a complete story of commercial risk.

And data that is not explicitly available can be surmised and generated through trained AI models. For instance, a carrier looking to inform employment practices liability insurance coverage asks in the underwriting forms whether a business has an employee handbook. This data point does not appear plainly over the web most of the time. However, trained machine models automatically realize that structured job postings with consistent business and position descriptions are indicative of an existing employee handbook. Another business might not explicitly state that they are regularly lifting and transporting heavy cargo, but AI image research could reveal the presence of forklifts on the premises.