In today’s fast-paced, digital marketplace, the value of pictures isn’t measured in thousands of words, but in millions of dollars. And in small commercial insurance, images can tell far more about a business than many questions on an application combined.
The answer to speedier and more accurate underwriting lies not in more time on research or more back-and-forth questions between underwriter and business owner, but in innovation.
Take a landscaper, for example, whose business activities may extend beyond cutting grass and spreading mulch. Photos of workers in action might reveal tree trimming high above the ground. Other images may show workers excavating a lawn to install drains. Either of these activities is critical for an underwriter to know about since they potentially create new exposures, but could easily be missed, especially if the work is intermittent.
Discovering the full range of a business’s activities and attributes that can impact premiums, coverage, and claims is critical for accurate underwriting but can be difficult in practice. Underwriters often rely on information available on websites and social media or provided by the insured or agent. Verifying information can be prohibitively time-consuming and may not produce a complete picture because data sources are fragmented and sometimes out of date. Agents are under increasing pressure to quote policies quickly, which can result in reliance on incomplete or even faulty applicant information.
In fact, a Verisk analysis found that just over half of small business policies were misclassified in Standard Industrial Classification/North American Industrial Classification System (SIC/NAICS) codes. Inaccurate classification could mean missing details of critical business attributes, which could impact both sides of the transaction, such as leaving the insured vulnerable if a policy expressly excludes an activity or an insurer collecting premiums misaligned with the risk.
As insurers strive to streamline and automate underwriting, spending effort on manual research is at odds with customer and agent expectations for a fast turnaround. The answer to speedier and more accurate underwriting lies not in more time on research or more back-and-forth questions between underwriter and business owner, but in innovation.
Image analytics emerges as an innovative underwriting tool
Images displayed on websites and social media can be a rich source of information about small businesses, providing detailed illustrations of business locations, snapshots uploaded by customers, and other images that can reveal a wide array of activities and possible exposures.
Advances in machine learning and artificial intelligence (AI) can turn content extracted from images into meaningful and practical underwriting insights, especially when combined with other robust data on the business. For example, photos of a landscaping business in action could include workers up in a cherry picker trimming branches. An image analytics model trained to “see” with an underwriter’s eye could identify a heavy machinery exposure. An insurer might use this insight to recommend heavy equipment insurance. A computer model could also confirm this business employs safety measures by having workers wear protective equipment, which would inform insurers they are taking steps to reduce risk and may be eligible for potential premium discounts.
And while humans can only interpret images one at a time, models powered by computer vision can automate this process to analyze multiple risks simultaneously.
Data and insurance expertise combined
To truly revolutionize an insurer’s workflow, image analytics technology must be paired with deep and broad insurance industry expertise to develop computer models that can identify insights relevant for underwriting. Photos of a restaurant, for instance, can reveal a lot of information, some little more than trivia that can distract from identifying details that can affect the risk profile. However, an image model that spots a bar and photos of a live band performing help an insurer know that they can provide appropriate coverage for liquor and event-related exposures.
Computer vision models that analyze images must not only be purpose-built for insurers but also thoroughly tested and evaluated for the accuracy of the insights they produce before any findings are released into production. Heavily vetting the results of AI-generated data can increase insurers’ confidence in the valuable business insights that image analytics models provide.
Delivering these insights at point-of-quote with just a business name and address has the power to transform the underwriting process by helping to answer—or even eliminate—most underwriting questions. Underwriters can straight-through process more of their small business risks, confident that actionable and reliable data fuels their decisions, driving both efficiency and profitability.
Discover how image analytics opens a new window to see your customers and applicants more clearly than ever before.