March 26, 2018 by David Gambrill
Artificial intelligence can now surgically identify Canadian homes and businesses in an insurer’s book that are most in need of inspection, thereby eliminating unnecessary home inspection orders.
“Sometimes the industry will pay to have an inspection done based on criteria and there’s nothing that we find wrong with the home, or anything that adds value to the insurer,” says Greg McCutcheon, president of Opta Information Intelligence, an SCM company.
“Using advanced machine learning techniques and our insights, we’ve created a predictive model that can predict [which] homes are dramatically underinsured, as well homes that have something undesirable from a risk perspective associated with them,” McCutcheon said. “I can identify accurately those homes that we know are underinsured by a factor of X, and [the insurer needs] to make sure they are properly insured to value.”
AI techniques are designed to create more value for insurers’ home inspection orders than the traditional scattershot approach. Currently, an insurer bases an inspection order on classical underwriting criteria such as the age of the home or replacement value. The problem is, if an insurer uses those static criteria to pay money to inspect a home or a business that is perfectly fine, the insurer has created absolutely no return on investment for the inspection order.
The traditional approach may be a factor in a worrisome trend for companies to emphasize new business inspections over inspecting renewal properties, McCutcheon says.
“A lot of people are ordering [inspections] on new business,” McCutcheon observes. “But they retain that property for up to eight or nine years. A lot can change on a home or a commercial business during that period, and it seems that we’ve stopped inspecting on renewals, but mostly inspect on new business. I think that’s a dangerous practice for the industry, because a lot can change.”
AI is a way to keep up with the changes. It’s also behind a new tool that can help insurers become more judicious in their choice of home inspection orders.
Currently available, the solution, called Inspection Score, uses AI to comb through satellite imagery and inspection data that Opta has amassed on more than 13 million homes and businesses across Canada. The solution benefited from Opta’s recent integration with its sister SCM company, Risk Management Services (RMS) — a move that essentially brought Opta’s data analytics capabilities and RMS’s field expertise in home inspections all under one roof.
Inspection Score uses AI to parse through a huge store of inspection data to look for signs of underinsured homes, picking out risks such as poor roof condition, improperly connected wood stoves, and undesirable wiring.
Some data comes from satellite imagery. Inspection Score also leverages data from iClarify, a tool that generates more than 30,000 updates a day through business transactions that are being processed across the country.
Based on all the information, the insurer gets a simple score, from 1 (low risk) to 100 (high risk), showing which properties in its book are best served by an inspection.